studies on the promiscuous subunit tina schwabeirep.ntu.ac.uk/id/eprint/197/1/203555_tinas...

Investigation of an unusual GABAA receptor -

Studies on the promiscuous ε subunit

Tina Schwabe

A thesis submitted in partial fulfilment of the requirements of

Nottingham Trent University for the degree of Doctor of Philosophy.

December 2010

Copyright statement

This work is the intellectual property of the author, and may also be owned by Nottingham

Trent University. You may copy up to 5 % of this work for private study, or personal, non-

commercial research. Any re-use of the information contained within this document should

be fully referenced, quoting the author, title, university, degree level and pagination. Queries

or requests for any other use, or if a more substantial copy is required, should be directed in

the first instance to the author.

i

Abstract

γ-aminobutyric acid (GABA) type A (GABAA) receptors are pentameric, chloride-selective

ion channels. Nineteen different GABAA receptor subunits are known to exist in man and va-

rious subunit combinations are able to form receptor subtypes with different, subunit-specific

pharmacological properties. GABAA receptors are targeted by many clinically-important

drugs, for example, benzodiazepines. Unwanted side effects, such as tolerance and depen-

dence, often occur due to non-selective receptor targeting. There is, therefore, a need for

subtype-selective drugs. Comparatively little is known about the role of the ε subunit or

ε-subunit-containing receptors in vivo. Recently, it has been recognised that a GABAA re-

ceptor subtype, that contains the α3 and ε subunits, is present in neurons within the forebrain

that synthesise acetylcholine, and this may regulate neurotransmitter release. In Alzheimer’s

disease, cholinergic neurotransmission is reduced; therefore, this subtype might be a target

for the development of anti-Alzheimer’s disease drugs that would function by increasing

acetylcholine release in forebrain regions where cholinergic neurons are dying. The phar-

macological properties of a GABAA receptor that comprises the α3, β2 and ε subunits were

examined in Xenopus laevis oocytes using the two-electrode voltage-clamp technique. The

results obtained confirm that the ε subunit confers unusual properties on the GABAA recep-

tor. α3β2ε receptors were spontaneously active and displayed high agonist sensitivity. Next,

a screening assay was established in human embryonic kidney 293 cells, which were tran-

siently transfected with complementary DNAs for GABAA receptor subunits and a yellow

fluorescence protein variant (YFP-H148Q/I152L). The latter is quenched by anion influx into

cells upon receptor activation. Using this assay, receptor activity and GABA-concentration

relationships could be demonstrated. It was, therefore, used to screen for compounds that

modulate the activity of ε-subunit-containing receptors. Several interesting compounds were

identified, and they were further examined using the two-electrode voltage-clamp technique.

These compounds are potential candidates for drug development, and may help to identify

the physiological role(s) of GABAA receptors that contain the ε polypeptide in vivo.

I would like to dedicate my thesis to my grandma Ursula Pöller.

Acknowledgements

I would like to thank my supervisor Professor Mark Darlison for giving me the opportunityto do my PhD in his laboratory, for the help and guidance I received from him and forintroducing me to the exciting field of ion channel research. I would also like to thank myother supervisors Professor John Wallis and Dr Chris Garner for their support.

I am grateful to my co-supervisor Dr Ian Mellor for all his help throughout my PhD and,especially, for giving me the opportunity to work in his laboratory at the University of Not-tingham on a regular basis. It has been a great experience and I very much enjoyed workingwith you. I would also like to thank Ian for proof-reading most parts of my thesis.

Next, I would like to thank all friends and colleagues at Nottingham Trent University, Uni-versity of Nottingham and Edinburgh Napier University for the help they have given meduring the various stages of my PhD. I very much appreciated the warm welcome, friendli-ness and support I received in Ians laboratory as well as in Edinburgh after I first arrived. Iam thankful to Amy Poole who has been a great help for me, especially during the last twoyears. It was nice to getting to know you better.

Most importantly, I would like to thank my family, especially my parents and grandparentsfor their constant support and encouragement. I very much missed seeing you more often.Finally, I would like to thank my fiancï£¡e Shiva Marthandan who has been there for methroughout these past few years, helping, encouraging and comforting me. Thanks for havingfaith in me and my work; this has greatly helped me to get to the end and finish my thesis.

Finally, I am thankful to Prof. Erwin Sigel (Institute of Biochemistry and Molecular Medi-cine, University of Bern, Switzerland) for supplying the GABAA receptor α3-, β2- and γ2L-subunit cDNAs, to Dr Maurice Garret (Laboratoire de Neurophysiologie, Bordeaux, France)for providing the ε-subunit cDNA, to Dr Thomas Kuner (Institute of Anatomy and Cell Bio-logy, Heidelberg University, Germany) for the pRK5-Clomeleon plasmid and to Prof. AlanVerkman (Department of Medicine, University of California, San Francisco, USA) for sup-plying the pEYFP H148Q/I152L mutant vector. I am also grateful to Prof. Anne Stephenson(The School of Pharmacy, University of London, UK) for providing HEK293 cells.

iv

Table of Contents

List of Figures viii

List of Tables xi

Abbreviations xiii

1 General Introduction 11.1 Alzheimer’s disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 GABA and GABAA receptors . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2.1 Molecular structure of GABAA receptors . . . . . . . . . . . . . . 31.2.2 GABAA receptor subunit diversity . . . . . . . . . . . . . . . . . . 41.2.3 GABAA receptor stoichiometry and distribution in the brain . . . . 61.2.4 Pharmacology of GABAA receptors . . . . . . . . . . . . . . . . . 81.2.5 Minor GABAA receptors . . . . . . . . . . . . . . . . . . . . . . . 13

1.2.5.1 The GABAA receptor ε subunit . . . . . . . . . . . . . . 141.3 Investigating ion channel properties . . . . . . . . . . . . . . . . . . . . . 19

1.3.1 Mammalian expression systems . . . . . . . . . . . . . . . . . . . 211.3.2 Xenopus laevis oocytes . . . . . . . . . . . . . . . . . . . . . . . . 231.3.3 Electrophysiological techniques . . . . . . . . . . . . . . . . . . . 251.3.4 Flux assays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271.3.5 Fluorescence-based assays . . . . . . . . . . . . . . . . . . . . . . 28

1.4 Aims and objectives of the study . . . . . . . . . . . . . . . . . . . . . . . 29

2 Materials and Methods 322.1 General molecular biology techniques . . . . . . . . . . . . . . . . . . . . 32

2.1.1 Preparation of competent Escherichia coli (E. coli) cells . . . . . . 322.1.2 Transformation of plasmid DNA into competent E. coli cells . . . . 332.1.3 Plasmid preparation . . . . . . . . . . . . . . . . . . . . . . . . . 33

2.1.3.1 Minipreparation . . . . . . . . . . . . . . . . . . . . . . 332.1.3.2 Maxipreparation . . . . . . . . . . . . . . . . . . . . . . 33

2.1.4 Restriction endonuclease digestion . . . . . . . . . . . . . . . . . . 342.1.5 Measuring nucleic acid concentrations . . . . . . . . . . . . . . . . 342.1.6 Ethanol precipitation . . . . . . . . . . . . . . . . . . . . . . . . . 352.1.7 Sequencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352.1.8 Side-directed mutagenesis - Generating a truncated ε subunit . . . . 35

2.1.8.1 Primer design . . . . . . . . . . . . . . . . . . . . . . . 352.1.8.2 Polymerase chain reaction . . . . . . . . . . . . . . . . . 372.1.8.3 Ligation of the εT subunit cDNA into the pcDNA3+ vector 39

2.2 GABAA receptor subunits & fluorescence proteins . . . . . . . . . . . . . . 40

v

TABLE OF CONTENTS

2.2.1 Recovery of vectors harbouring GABAA receptor subunits . . . . . 402.2.2 Recovery of pRK5-Clomeleon and pEYFP H148Q/I152L vectors . 402.2.3 Examination of received vectors . . . . . . . . . . . . . . . . . . . 412.2.4 Subcloning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.3 Electrophysiological recordings in Xenopus oocytes . . . . . . . . . . . . . 422.3.1 Making cRNA from cDNA . . . . . . . . . . . . . . . . . . . . . . 422.3.2 Selection of Xenopus laevis oocytes . . . . . . . . . . . . . . . . . 442.3.3 Injection of cRNA into Xenopus laevis oocytes . . . . . . . . . . . 452.3.4 Two-electrode voltage-clamp . . . . . . . . . . . . . . . . . . . . . 452.3.5 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

2.4 Patch clamping mammalian cells . . . . . . . . . . . . . . . . . . . . . . . 472.5 Cell culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

2.5.1 Culturing HEK293 cells . . . . . . . . . . . . . . . . . . . . . . . 482.5.2 Long term storage of HEK293 cells in liquid nitrogen . . . . . . . . 492.5.3 Thawing cells stored in liquid nitrogen . . . . . . . . . . . . . . . 502.5.4 Transient transfection of HEK293 cells . . . . . . . . . . . . . . . 50

2.5.4.1 Evaluation of transient transfection efficiency . . . . . . 512.5.5 Reseeding cells into appropriate tissue culture plasticware . . . . . 51

2.6 Assay development to screen compound libraries . . . . . . . . . . . . . . 522.6.1 A cell-based assay using the chloride-ion indicator Clomeleon . . . 52

2.6.1.1 Excitation and emission spectra of Clomeleon . . . . . . 522.6.1.2 Evaluating a screening assay using Clomeleon . . . . . . 52

2.6.2 Setting up an iodide-flux assay . . . . . . . . . . . . . . . . . . . . 532.6.3 Setup and optimisation of a mutant YFP-based assay . . . . . . . . 55

2.6.3.1 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . 572.7 Compound screening using mutant YFP-H148Q/I152L . . . . . . . . . . . 58

2.7.1 Chemical compound library . . . . . . . . . . . . . . . . . . . . . 582.7.2 Screening assay . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

2.7.2.1 Data analysis . . . . . . . . . . . . . . . . . . . . . . . . 592.7.3 Further testing of ’hit’ compounds . . . . . . . . . . . . . . . . . . 59

3 Pharmacological characterisation of the α3β2ε GABAA receptor subtype 613.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.1.1 The GABAA receptor ε subunit . . . . . . . . . . . . . . . . . . . 613.1.2 ε subunit splice variants . . . . . . . . . . . . . . . . . . . . . . . 62

3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 663.2.1 Agonist sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . . 663.2.2 Current-voltage relationships . . . . . . . . . . . . . . . . . . . . . 683.2.3 Effect of zinc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 683.2.4 Application of the benzodiazepine flunitrazepam . . . . . . . . . . 693.2.5 Constitutive activity of the α3β2ε GABAA receptor subtype . . . . 703.2.6 Effect of etomidate on α3β2ε GABAA receptors . . . . . . . . . . 713.2.7 Expression of a truncated ε splice variant . . . . . . . . . . . . . . 74

3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753.3.1 Pharmacological characteristics of the ε subunit . . . . . . . . . . . 753.3.2 Investigation of a truncated ε subunit variant . . . . . . . . . . . . 793.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

vi

TABLE OF CONTENTS

4 Development of an assay to screen compound libraries 814.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

4.1.1 Clomeleon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 824.1.2 Iodide-flux method . . . . . . . . . . . . . . . . . . . . . . . . . . 844.1.3 YFP-based assays . . . . . . . . . . . . . . . . . . . . . . . . . . . 87

4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 894.2.1 Evaluating a screening assay using Clomeleon . . . . . . . . . . . 894.2.2 Iodide-flux assay . . . . . . . . . . . . . . . . . . . . . . . . . . . 934.2.3 Optimisation of a YFP-based screening assay . . . . . . . . . . . . 103

4.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.3.1 Clomeleon-based screening assay . . . . . . . . . . . . . . . . . . 1084.3.2 Colorimetric iodide-flux assay . . . . . . . . . . . . . . . . . . . . 1104.3.3 YFP-based screening assay . . . . . . . . . . . . . . . . . . . . . . 1154.3.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

5 Screening a chemical compound library 1195.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1195.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

5.2.1 Preliminary screening . . . . . . . . . . . . . . . . . . . . . . . . 1215.2.2 Investigation of ’hit’ compounds on α1β2ε receptors . . . . . . . . 1285.2.3 Investigation of ’hit’ compounds on α3β2γ2 receptors . . . . . . . 1495.2.4 Investigation of ’hit’ compounds in Xenopus laevis oocytes . . . . . 158

5.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

6 General Discussion 1676.1 Pharmacological characterisation of the α3β2ε GABAA receptor . . . . . . 1676.2 Search for and setup of a cell-based screening assay . . . . . . . . . . . . . 1696.3 Screening of α1β2ε and α3β2γ2 GABAA receptors . . . . . . . . . . . . . 1716.4 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1726.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173

Bibliography 175

Appendices 191

A Vector maps 192

B Additional Graphs 195

vii

List of Figures

1.1 GABAA receptor structure . . . . . . . . . . . . . . . . . . . . . . . . . . 41.2 Schematic illustration of the GABAA receptor with some drug binding sites 81.3 Genetic engineering studies revealed GABAA receptor subunit functions . . 101.4 Possible subunit arrangements in ε-containing GABAA receptors . . . . . . 161.5 Interrelations between different attributes of ion channel assays . . . . . . . 201.6 Xenopus laevis oocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231.7 Schematic diagram of the functional expression of genetic information in

Xenopus oocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241.8 Schematic representation of the two-electrode voltage-clamp technique . . 261.9 Schematic representation of the patch-clamp whole-cell configuration . . . 27

2.1 Side-directed mutagenesis using overlap PCR to create εT . . . . . . . . . . 362.2 Healthy and impaired Xenopus laevis oocytes . . . . . . . . . . . . . . . . 44

3.1 Alignment of the partial rat GABAA receptor ε subunit amino acid sequence 643.2 Structure of the ε subunit . . . . . . . . . . . . . . . . . . . . . . . . . . . 653.3 Sample current traces evoked by ascending GABA concentrations in oocytes

expressing different receptor subtypes . . . . . . . . . . . . . . . . . . . . 663.4 Effect of GABA on receptors expressed in Xenopus oocytes . . . . . . . . . 673.5 Muscimol activation of GABAA receptors expressed in oocytes . . . . . . . 673.6 Current-voltage relationships of GABA-evoked currents in Xenopus oocytes 683.7 Effect of zinc on GABAA receptors expressed in Xenopus oocytes . . . . . 693.8 Effect of flunitrazepam on GABAA receptors expressed in oocytes . . . . . 703.9 Current traces evoked by GABA and picrotoxin in oocytes expressing α3β2

or α3β2ε receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 713.10 Effect of etomidate on oocytes expressing α3β2 and α3β2ε receptors . . . 733.11 Effect of GABA and zinc on oocytes expressing α3β2εT receptors . . . . . 74

4.1 Principle of Clomeleon, a FRET-based chloride-indicator protein . . . . . . 834.2 Assay principle of the colorimetric iodide-flux assay . . . . . . . . . . . . 864.3 Schematic diagram showing the principle of the YFP screening assay . . . 884.4 Emission scan of the chloride-indicator protein Clomeleon . . . . . . . . . 904.5 Excitation scan of the chloride-indicator protein Clomeleon . . . . . . . . . 914.6 Fluorescence signal of Clomeleon in HEK293 cells expressing GABAA re-

ceptors after addition of the agonist GABA . . . . . . . . . . . . . . . . . 924.7 NaI flux assay standard curve . . . . . . . . . . . . . . . . . . . . . . . . . 944.8 OD405 readings obtained from HEK293 cells expressing GABAA receptors 954.9 Iodide efflux from HEK293 cells using SK assay . . . . . . . . . . . . . . 974.10 OD405 values of cells exposed to GABA with different incubation times . . 98

viii

LIST OF FIGURES

4.11 Iodide-flux assay results after applying GABA over different time periods . 994.12 SK assay results obtained from different HEK293 cell densities . . . . . . . 1014.13 SK assay applying GABA, bicuculline and picrotoxin to the cells . . . . . . 1024.14 Current traces evoked by GABA and picrotoxin in HEK293 cells expressing

α3β2γ2 receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1034.15 Net fluorescence signal of mutant YFP-H148Q/I152L in HEK293 cells ex-

pressing α3-subunit-containing GABAA receptors . . . . . . . . . . . . . . 1054.16 Net fluorescence signal of mutant YFP-H148Q/I152L in HEK293 cells ex-

pressing α1-subunit-containing GABAA receptors . . . . . . . . . . . . . . 1064.17 GABA concentration-response relations in a YFP-based screening assay . . 1074.18 Fluorescence signal change after addition of flunitrazepam to cells expres-

sing α3β2γ2 GABAA receptors . . . . . . . . . . . . . . . . . . . . . . . 108

5.1 Sample negative results from preliminary screen of α1β2ε receptors . . . . 1225.2 Sample negative results from preliminary screen of α3β2γ2 receptors . . . 1235.3 Representative positive screening results for α1β2ε GABAA receptors . . . 1245.4 Representative positive screening results for α3β2γ2 GABAA receptors . . 1255.5 Results obtained for different concentrations of a test compound in cells ex-

pressing α1β2ε receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . 1275.6 Results obtained for different concentrations of a test compound in cells ex-

pressing α3β2γ2 receptors . . . . . . . . . . . . . . . . . . . . . . . . . . 1275.7 Effect of compound TS-1 on α1β2ε-receptor-containing cells . . . . . . . . 1295.8 Effect of compound TS-2 on cells expressing α1β2ε GABAA receptors . . 1305.9 Effect of compound TS-3 on cells expressing α1β2ε GABAA receptors . . 1315.10 Effect of compound TS-4 on cells expressing α1β2ε GABAA receptors . . 1325.11 Effect of compound TS-5 on cells expressing α1β2ε GABAA receptors . . 1335.12 Effect of compound TS-6 on cells expressing α1β2ε GABAA receptors . . 1345.13 Effect of compound TS-7 on α1β2ε-receptor-containing cells . . . . . . . . 1355.14 Effect of compound TS-8 on α1β2ε-receptor-containing cells . . . . . . . . 1365.15 Effect of compound TS-9 on cells expressing α1β2ε GABAA receptors . . 1375.16 Effect of compound TS-10 on cells expressing α1β2ε GABAA receptors . . 1385.17 Effect of compound TS-11 on cells expressing α1β2ε GABAA receptors . . 1395.18 Effect of compound TS-12 on cells expressing α1β2ε GABAA receptors . . 1405.19 Effect of compound TS-13 on cells expressing α1β2ε GABAA receptors . . 1415.20 Effect of compound TS-14 on cells expressing α1β2ε GABAA receptors . . 1425.21 Effect of compound TS-15 on cells expressing α1β2ε GABAA receptors . . 1435.22 Effect of compound TS-16 on cells expressing α1β2ε GABAA receptors . . 1455.23 Effect of compound TS-17 on cells expressing α1β2ε GABAA receptors . . 1465.24 Effect of compound TS-18 on cells expressing α1β2ε GABAA receptors . . 1475.25 Effect of compound TS-19 on α1β2ε-receptor-expressing cells . . . . . . . 1485.26 Effect of compound TS-10 on cells expressing α3β2γ2 GABAA receptors . 1505.27 Effect of compound TS-20 on cells expressing α3β2γ2 GABAA receptors . 1515.28 Effect of compound TS-21 on cells expressing α3β2γ2 GABAA receptors . 1525.29 Effect of compound TS-22 on cells expressing α3β2γ2 GABAA receptors . 1535.30 Effect of compound TS-23 on cells expressing α3β2γ2 GABAA receptors . 1555.31 Effect of compound TS-16 on cells expressing α3β2γ2 GABAA receptors . 1565.32 Effect of compound TS-24 on cells expressing α3β2γ2 GABAA receptors . 1575.33 Effect of compound TS-3 on oocytes expressing α1β2ε GABAA receptors . 159

ix

LIST OF FIGURES

5.34 Effect of compound TS-16 on oocytes expressing α1β2ε or α3β2γ2 receptors 1605.35 Effect of compound TS-23 on oocytes expressing α3β2γ2 GABAA receptors 161

A.1 pcDNA3.1 vector map . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192A.2 pcDNA3 vector map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193A.3 pCMV vector map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193A.4 pBS vector map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194A.5 pRK5-Clomeleon vector map . . . . . . . . . . . . . . . . . . . . . . . . . 194

B.1 Effect of DMSO on oocytes expressing different GABAA receptor subtypes 195B.2 Fluorescence signal of mutant YFP-H148Q/I152L in response to different

concentrations of DMSO . . . . . . . . . . . . . . . . . . . . . . . . . . . 196B.3 OD405 values of cells exposed to GABA or DPBS with different incubation

times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197B.4 OD405 values of cells exposed to GABA or DPBS with different incubation

times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198B.5 Effect of GABA on α1β2ε receptors expressed in Xenopus oocytes . . . . . 199B.6 Effect of GABA on α3β2γ2 receptors expressed in HEK293 cells . . . . . 199

x

List of Tables

1.1 GABAA receptor subunit clusters in the human genome . . . . . . . . . . . 61.2 Native GABAA receptors in the brain . . . . . . . . . . . . . . . . . . . . . 7

2.1 Oligonucleotide design to generate a GABAA receptor εT subunit . . . . . . 352.2 Restriction enzymes used to digest different GABAA receptor cDNAs . . . 412.3 Transient transfection using TransPassTM D2 in different culture vessels . . 50

3.1 Effects of GABA and muscimol on different GABAA receptors . . . . . . . 67

4.1 Analysis of OD405 readings obtained from SK assay using HEK293 cells . 95

xi

Abbreviations

5-HT3Rs Serotonin 5-hydroxytryptamine3 receptorsA AbsorbanceACh AcetylcholineAD Alzheimer’s diseaseAE1 Anion exchanger 1BF Baseline fluorescenceBLAST Basic local alignment search toolbp Base pairscDNA Complementary DNACFP Cyan fluorescence proteinCFTR Cystic fibrosis transmembrane conductance regulator proteinCNS Central nervous systemcRNA Complementary RNADEPC DiethylpyrocarbonateDMCM methyl-6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylateDMEM Dulbecco’s modified Eagle’s mediumDMSO Dimethyl sulfoxideDPBS Dulbecco’s phosphate buffered salineE.coli Escherichia coliεT Truncated ε subunitEC50 Effective concentration for 50 % receptor activityEDTA Ethylenediaminetetraacetic acidFBS Fetal bovine serumFLIPR Fluorometric Imaging Plate ReaderFRET Förster (or Fluorescence) Resonance Energy TransferGABA γ-Aminobutyric acidGABAA GABA type AGABAB GABA type BGAD Glutamate decarboxylaseGFP Green fluorescence proteinGlyRs Glycine receptorsHEK Human embryonic kidneyHTS High throughput screeningIC50 Effective concentration for 50 % receptor inhibitionKApp Apparent dissociation constantKd Dissociation constantkb Kilo basesLB Lysogeny brothLC Locus coeruleus

xii

ABBREVIATIONS

LGICs Ligand-gated ion channelsMQAE N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromideMRK-016 3-tert-butyl-7-(5-methylisoxazol-3-yl)-2-(1-methyl-1H-1,2,4-triazol-5-

ylmethoxy)-pyrazolo[1,5-d]-[1,2,4]triazinemRNA Messenger RNAnAChRs Nicotinic ACh receptorsNKCC Na+-K+-Cl- co-transportersOD Optical densityPCR Polymerase chain reactionPKC Protein kinase Crpm Revolutions per minuterTEV Recombinant Tobacco Etch VirusSD Standard deviationSEM Standard error of the meanSK Sandell-KolthoffTm Melting temperatureTAE Tris-acetateTM Transmembrane domainUV Ultravioletv/v Volume per volumew/v Weight per volumeYFP Yellow fluorescence proteinYFP-H148Q/I152L YFP with mutation at positions 148 (histidine to glutamine) and 152

(isoleucine to leucine)

xiii

Chapter 1

General Introduction

1.1 Alzheimer’s disease

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder and the most frequent

form of dementia in elderly people. More than 25 million people worldwide are suffering

from dementia, a number that is expected to double every 20 years as the general age of

the human population increases (Ferri et al., 2005; Citron, 2010). Clinical features of AD

include loss of memory, cognitive dysfunction as well as behavioural disturbances, such as

mood disorders, agitation and psychosis, leading to death on average nine years after AD

was first diagnosed. AD is the sixth leading cause of death in the USA. The illness not

only has a dramatic effect on the patient, but there is also a considerable burden on close

relatives and caregivers, and ultimately national economies due to the frequent requirement

for institutionalisation. Total costs caused by AD worldwide are thought to be more than 200

billion dollars per annum (Citron, 2010).

The pathological hallmarks of AD are the presence of extracellular plaques containing β-

amyloid peptides, which derive from the amyloid precursor protein, and neurofibrillary

tangles, which contain abnormally phosphorylated forms of the cytoskeleton-associated pro-

tein tau, as well as the degeneration of synapses and neurons (Kar et al., 2004; Citron, 2010).

It is also well established that there is a loss of cholinergic neurotransmission in the hippo-

campus and neocortex of the brain. This is the result of reduced choline uptake, acetylcholine

1

Chapter 1. General Introduction

(ACh) release and choline acetyltransferase activity and, ultimately, the loss of basal fore-

brain cholinergic neurons. ACh is essential for cell-to-cell communication in the forebrain,

and neuronal pathways regulated by this transmitter are important for cognition. The fin-

dings outlined above have led to the ’cholinergic hypothesis’ of AD, in which the loss of

basal forebrain cholinergic neurons and the interruption of cholinergic neurotransmission

in the cerebral cortex and other areas is proposed to cause the observed cognitive impair-

ments (Kar et al., 2004; Yuede et al., 2007). Currently, there is no cure for this debilitating

illness. However, therapeutic approaches have been developed, for example ACh-esterase

inhibitors, which increase ACh levels in the brain resulting in improved cognitive functions.

There are currently four approved ACh-esterase inhibitors to treat mild to moderate AD: ta-

crine (Cognex®), donepezil (Aricept®), rivastigmine (Exelon®) and galantamine (Reminyl®;

Yuede et al., 2007). The latter also potentiates the activity of neuronal nicotinic ACh re-

ceptors (nAChRs), which is believed to further improve neurotransmission (Seltzer, 2010).

However, these drugs merely delay the progression of disease symptoms for one to two years.

Moreover, the ACh-esterase inhibitor tacrine has been shown to trigger hepatotoxic effects

and its use has been diminished (Watkins et al., 1994). Hence, to treat AD more efficiently,

new drug targets and more effective treatments are urgently needed.

1.2 GABA and GABAA receptors

γ-Aminobutyric acid (GABA) is a four-carbon nonprotein amino acid. It is synthesised from

L-glutamic acid catalysed by glutamate decarboxylase (GAD); GAD requires the cofactor

pyridoxal phosphate (Martin & Rimvall, 1993). It exists in two isoforms, GAD65 (65 kDa

isoform) and GAD67 (67 kDa isoform). The two forms are regulated separately and localised

in different subcellular compartments. GAD67 is present throughout the cytoplasm of neu-

rons, generating GABA for metabolic purposes rather than neurotransmission, for example

protection after neuronal injury or regulation of redox potential in response to oxidative stress

(Kaufman et al., 1991; Waagepetersen et al., 1999; Lamigeon et al., 2001; Buddhala et al.,

2009). GAD65 is present in nerve terminal endings generating GABA for neurotransmission

purposes. GABA is removed from the synaptic cleft by high-affinity transporters and bro-

2


ken down by GABA transaminase into succinic semialdehyde and glutamic acid (D’Hulst &

Kooy, 2007).

GABA is the main inhibitory neurotransmitter in the mammalian central nervous system

(CNS) and found in 30 % of neuronal synapses. Its effects are mediated via ionotropic GABA

type A (GABAA) receptors or via metabotropic GABA type B (GABAB) receptors. The latter

are coupled to G proteins, which mediate signal transduction via secondary messengers,

whereas the former are directly gating Cl- ions (Bormann, 2000). GABAA receptors are

widespread throughout the mammalian CNS where they control the majority of physiological

actions of GABA, including learning and memory processes, cognition, feeding and drinking

behaviour and the modulation of anxiety (Bormann, 2000; Sieghart & Sperk, 2002; D’Hulst

et al., 2009). Mutations or functional deficits of GABAA receptors have been implicated with

a variety of neuropsychiatric diseases, such as anxiety, schizophrenia, fragile X syndrome

and several genetic forms of epilepsy (D’Hulst & Kooy, 2007; Moult, 2009; Macdonald

et al., 2010). The receptors importance is further reflected by the fact that it is targeted by

a plethora of pharmacologically and clinically important drugs, including benzodiazepines

and anaesthetics.

1.2.1 Molecular structure of GABAA receptors

GABAA receptors belong to the superfamily of Cys-loop ligand-gated ion channels (LGICs)

that also includes nAChRs, serotonin 5-hydroxytryptamine3 receptors (5-HT3Rs), glycine

receptors (GlyRs) and zinc-activated ion channels (Schofield et al., 1987; Olsen & Sie-

ghart, 2008). Cys-loop LGICs mediate fast synaptic neurotransmission by conducting ca-

tions (nAChRs, 5-HT3Rs and zinc-activated ion channels) or anions (GABAA receptors and

GlyRs), mediating excitatory or inhibitory neurotransmission, respectively (Pless & Lynch,

2008). GABAA receptors consist of five subunits, each of which comprises a large extracel-

lular amino terminal domain of about 220 amino acids, four hydrophobic, α-helical trans-

membrane domains (TM1-4) each of about 20 amino acids in length, a large intracellular

loop between TM3 and TM4 and a short extracellular carboxyl-terminal domain (Figure 1.1

A). The extracellular amino terminal domain includes a 13 amino acid sequence flanked by

3


two cystein residues that form a disulphide bond. This characteristic structural feature, com-

mon to all Cys-loop LGICs, determined the superfamily’s name (Simon et al., 2004). The

receptor subunits are arranged in a circle to form a central ion-conducting channel, with the

second TM of each subunit forming the channel pore (Figure 1.1 B; Hevers & Lüddens,

1998). Residues within the amino terminal domain of two adjacent subunits form the ligand

binding sites. Binding of two GABA molecules to a GABAA receptor triggers a conformatio-

nal change in the receptor, resulting in the opening of the intrinsic channel pore, the influx of

extracellular ions (primarily Cl- ions) via diffusion and subsequent hyperpolarisation of the

post-synaptic membrane, which leads to an increase in the inhibitory tone (Darlison et al.,

2005).

N

C

TM

1

TM

2

TM

3

TM

4

intracellular

extracellular

A B

Figure 1.1: GABAA receptor structure. (A) Schematic representation of a receptor subunitand (B) the pentameric receptor structure.

1.2.2 GABAA receptor subunit diversity

Nineteen different GABAA receptor subunits are known to exist in man; analysis of the

human genome database revealed that there are no further GABAA receptor subunit genes

(Simon et al., 2004). The subunit genes are subdivided into classes according to their amino

acid sequence homology. Sequence similarities between members of the same family are

around 70 % and between different families around 20 to 40 % (Darlison et al., 2005). There

are six alpha (α1-α6), three beta (β1-β3), three gamma (γ1-γ3), one delta (δ), one epsilon

4


(ε), one theta (θ) and one pi (π) subunit as well as three rho (ρ1-ρ3) subunits. The latter

were formerly classified as GABA type C receptors due to their distinct pharmacological

characteristics and the fact that they are highly expressed in the vertebrate retina and do

not co-localise with other GABAA receptor subtypes (Bormann, 2000; Olsen & Sieghart,

2008). A number of receptor subunits exhibit splice variants, among others the α6, γ2 and ε

subunits, generating further structural diversity of GABAA receptors (see also Section 3.1.2,

Chapter 3; Whiting et al., 1990, 1997; Korpi et al., 1994; Simon et al., 2004). Furthermore,

two other subunits, β4 and γ4, have been identified in lower vertebrates which lack the ε

and θ subunits (Bateson et al., 1991; Harvey et al., 1993). Based on sequence similarities

between the ε and γ4 subunits (49 % identity) as well as between the θ and β4 subunits

(56 % identity) and gene mapping studies, it has been concluded that the ε and θ subunits

are orthologues of the γ4 and β4 subunits, respectively (Simon et al., 2004; Darlison et al.,

2005).

Each receptor subunit is encoded by a distinct gene; most of which are clustered on dif-

ferent chromosomes in the human genome (see Table 1.1; Simon et al., 2004; Darlison et al.,

2005). The subunit genes are expressed independently from other genes of the same clus-

ter and associate together in different combinations to form a variety of receptor subtypes.

This explains the large number of different GABAA receptor subtypes present in the verte-

brate brain (see below). However, co-regulation of clustered genes may play a role in some

receptor subtypes, for example in α1β2γ2 receptors. All three subunits are clustered on

chromosome 5; the β2 and α1 genes are separated by the α6 subunit gene (β2-α6-α1-γ2).

It was shown in α6 ’knockout’ mice that α1 and β2 subunits are most likely regulated to-

gether by a common regulatory element, since expression of both subunits was reduced in

mice lacking the α6 subunit (Uusi-Oukari et al., 2000). Furthermore, α1 and β2 subunits are

widely expressed and exhibit a nearly identical expression pattern in the mammalian brain.

Expression of the γ2 subunit was not affected in α6 ’knockout’ mice compared to wild-type

mice. Nevertheless, these results suggest that subunit gene-clustering may partly contribute

to α1β2γ2 receptor assembling and may also play a role in subunit assembling of certain

other receptor subtypes.

5


Table 1.1: GABAA receptor subunit clusters in the human genome

GABAA receptor subunits Chromosomeα2, α4, β1, γ1 4p12-p13α1, α6, β2, γ2 5q31-q35α5, β3, γ3 15q11-q12α3, θ, ε Xq28ρ1, ρ2 6q14-q21

The π subunit gene is located on chromosome 5 (5q34-q35), but apart from the β2-α6-α1-γ2 subunitgene cluster. The δ subunit is present separately from all other GABAA receptor genes on chromo-some 1 (1p36.3; Simon et al., 2004; Darlison et al., 2005).

1.2.3 GABAA receptor stoichiometry and distribution in the brain

Hypothetically, if all existing GABAA receptor subunits could co-assemble at random, the

number of potential receptor subtypes would be huge. However, cell type-specific regulation

mechanisms control subcellular localisation, trafficking and subunit assembly (Fritschy &

Brünig, 2003; Lüscher & Keller, 2004), resulting in the existence of only a fraction of those

potential GABAA receptors in vivo. Sieghart (2000) estimated that there could be more than

500 different receptor subtypes expressed in the brain. Although, the presence of α and β

subunits is sufficient to form functional GABA-gated channels, most of the native receptors

are composed of three different types of subunits, usually two α subunits, two β subunits and

one γ subunit with a common stoichiometry of α-β-α-β-γ (Baumann et al., 2002; Sieghart

& Sperk, 2002). The γ subunit is in certain receptor subtypes replaced by a δ or ε subunit.

The distribution of the different GABAA receptor subtypes is brain area and cell type specific;

distinct neurons can express more than one receptor subtype. Subunit co-localisation in

different brain areas is important for receptor subunit composition. In situ hybridisation

and immunohistochemistry experiments revealed that α1, β1, β2, β3 and γ2 subunits are

widely expressed throughout the brain, albeit with different distribution patterns (Sieghart

& Sperk, 2002; Olsen & Sieghart, 2009). Expression of α2, α3, α4, α5, α6, γ1 and δ

subunits is less abundant and restricted to specific brain areas. The α2 subunit, for example,

is predominately expressed in the forebrain, whereas the α6 subunit is located only in granule

cells of the cerebellum and the cochlear nucleus (Pirker et al., 2000). Furthermore, the

majority of serotonergic neurons located in the raphne nuclei contain α3 subunits, but lack

6


α1 subunits (Gao et al., 1993). In contrast, GABAergic neurons in this area express both α

subunits and, hence, contain different receptor types.

The most common receptor subtype is the α1β2γ2 receptor, which is present in most areas

of the brain and makes up for 60 % of all GABAA receptors (Möhler, 2006, 2007). Other

common receptor subtypes are α2β3γ2 and α3βγ2 receptors, which count for 15 to 20 %

and 10 to 15 % of all receptors, respectively. Despite extensive research in the area, only

ten different native receptor subtypes have been identified indubitably by taking into account

specific criteria for receptor identification, including localisation, subunit composition, stoi-

chiometry and receptor pharmacology (see Olsen & Sieghart, 2008, 2009). Further receptor

subtypes have been proposed to exist with high probability or tentatively (see Table 1.2).

Table 1.2: Native GABAA receptors in the brain

Identified Existing with high probability Tentativeα1β2γ2 α1β3γ2 ρ1α2βγ2 α1βδ ρ2α3βγ2 α5β3γ2 ρ3α4βγ2 αβ1γ/αβ1δ αβγ1α4β2δ αβ αβγ3α4β3δ α1α6βγ/α1α6βδ αβθα5βγ2 αβεα6βγ2 αβπα6βδ αxαyβγ2ρ

By investigating different aspects of GABAA receptors, including localisation, subunit compositionand receptor pharmacology, ten native receptor subtypes have been unequivocally identified, furtherreceptor subtypes exist with high probability and others tentatively (adapted from Olsen & Sieghart,2009).

Although most receptors contain only one type of α or β subunits several studies have sug-

gested that there may be receptor subtypes co-expressing two different α or β subunits in

the brain (Sieghart & Sperk, 2002; Benke et al., 2004; Olsen & Sieghart, 2009). Immuno-

precipitation and -purification experiments and expression of α1, α6, β2 and γ2 subunits in

Xenopus laevis oocytes suggested that α1 and α6 subunits can co-exist in α1α6β2γ2 recep-

tor subtypes (Jechlinger et al., 1998; Sigel & Baur, 2000). Similarly, Fisher & Macdonald

(1997) showed that two different β subunits can co-exist in recombinant GABAA receptors

7


expressed in L929 fibroblasts conferring different properties to the receptor compared to

receptors containing only one of the two β subunits.

Different receptor subtypes exhibit distinct functional, physiological and pharmacological

characteristics. The specific subunit composition as well as the subunits position in the

pentameric receptor structure determine the receptors affinity for GABA, its pharmacological

profile as well as the receptors assembling patterns and targeting to different subcellular

domains (Ebert et al., 1994; Connolly et al., 1996; Whiting et al., 1999).

1.2.4 Pharmacology of GABAA receptors

As mentioned above, GABAA receptors have two GABA binding sites. They are located at

the interface between the α and β subunits and amino acid residues on both subunits form

the agonist binding pocket (Figure 1.2; Sigel & Buhr, 1997). Additionally, GABAA recep-

tors possess multiple other binding sites for a variety of pharmacologically and clinically

important drugs, such as benzodiazepines, general anaesthetics, neurosteroids, barbiturates

and convulsants (Figure 1.2 A; Bateson, 2004).

βα

αγ

Cl-

β

Benzodiazepine

GABAGABA

PTXZincEthanol

αααα

αααα

ββββ

ββββ

γγγγ Cl-

A B

Loreclezole/ etomidate

Figure 1.2: (A) Schematic illustration of the GABAA receptor and some of the several drugbinding sites. (B) Cross section of the pentameric receptor structure. Yellow circles representthe agonist GABA, which binds to the receptor between the α and β subunits. The bindingsite for benzodiazepines (red circles) is located at the interface of α and γ subunits.

8


The benzodiazepine binding site has received notable interest due to the clinical impor-

tance of benzodiazepines. Full agonists, such as diazepam, antagonists, such as methyl-

6,7-dimethoxy-4-ethyl-beta-carboline-3-carboxylate (DMCM), or inverse agonists, such as

flumazenil, but no endogenous ligands have been identified to bind to the site (van Niel et al.,

2005). Full benzodiazepine agonists have been used since the 1960s in a wide range of cli-

nical applications, such as treatment of anxiety disorders and insomnia and as anti-epileptic

and muscle relaxing agents (Bateson, 2004; Rudolph & Möhler, 2006). Inverse agonists

have been used as research tools only, due to their proconvulsant effects (van Niel et al.,

2005). The allosteric binding site is located at the interface between the α and γ subunits

(Pritchett et al., 1989b; Stephenson et al., 1990) and benzodiazepine ligand pharmacology

is determined by both the α and γ subunits of a receptor subtype (Pritchett et al., 1989a;

Hadingham et al., 1993; Wafford et al., 1993). Amino acid residues of the benzodiazepine

binding pocket are homologues to those in the GABA binding site (Olsen et al., 2004). It has

been shown that the γ2 subunit confers high sensitivity to classical benzodiazepines, such

as diazepam and triazolam, whereas receptors containing γ1 or γ3 subunits show decreased

affinity (Knoflach et al., 1991; Wafford et al., 1993). The α1, α2, α3 and α5 subunits also

confer sensitivity to classical benzodiazepines (Pritchett et al., 1989a; Smith & Olsen, 1995;

Sigel & Buhr, 1997), whereas receptors containing the α4 and α6 subunits are insensitive

to most clinically used benzodiazepines (Wisden et al., 1991; Möhler et al., 2002). Recep-

tor subtypes containing the δ and ε subunits are also benzodiazepine insensitive (Saxena &

Macdonald, 1994; Whiting et al., 1997).

Work with genetically engineered mice and studies using subunit specific ligands revealed

that the various effects of benzodiazepines (sedative/hypnotic, anxiolytic, anticonvulsant,

muscle relaxing and amnesic) are mediated by different GABAA receptor subtypes (McKer-

nan & Whiting, 1996; Rudolph et al., 1999; Rudolph & Möhler, 2006). In these mice, point-

mutations were introduced in α1, α2, α3 or α5 subunits, which caused insensitivity of the

specific subunit to diazepam (Rudolph & Möhler, 2004). Deficits in the behaviour of the

mutant mice were then examined and revealed the actions mediated by the mutated subunit

(Figure 1.3). By this means, it was demonstrated that receptors containing the α1 subu-

9


nit mediate the sedative, amnesic as well as partly the anticonvulsant effects of diazepam

(Rudolph et al., 1999). In contrast, the anxiolytic and myorelaxing effects of diazepam are

mediated by receptors containing the α2 subunit (Löw et al., 2000; Rudolph et al., 2001). It

was further revealed that α3 subunits mediate myorelaxation as well as an anxyolytic effect

at high receptor occupation. Several studies investigating α5 subunit functions revealed a

role in learning and memory, for example, mice lacking the α5 subunit performed better in a

spatial learning exercise than wild type mice (Rudolph & Möhler, 2006).

ββββ3ββββ2

αααα5αααα3αααα2αααα1

Diazepam

Sedation

Anterograde amnesia

Anticonvulsion(partly)

Anxiolysis Myorelaxation

Learning, memory

Etomidate

Hypnosis Immobility

α1/2/3/5 mutant mice

β2/3 mutant mice

Figure 1.3: Genetic engineering studies in mice revealed distinct actions of diazepam andetomidate on different GABAA receptor subtypes.Point-mutations in specific α or β subunits rendered them insensitive to diazepam or etomidate,respectively. Behavioural deficits in these genetically engineered mice gave information about theactions mediated by these specific receptor subtypes. For example, α1-containing receptors are me-diating the sedative action of benzodiazepines (adapted from Agid et al., 2007).

10


Subunit specific ligands have been shown to also demonstrate GABAA receptor subunit func-

tions and could confirm results obtained from genetically engineered mice (Chambers et al.,

2004; Sternfeld et al., 2004; Atack et al., 2006; Rudolph & Möhler, 2006). For example,

compound MRK-016 is an inverse agonist on the benzodiazepine binding site and acts pri-

marily on α5-subunit-containing receptors mediating enhanced cognition (Chambers et al.,

2004; Atack et al., 2009). Another example is the hypnotic, non-benzodiazepine zolpidem.

It binds preferentially to receptors containing the α1 subunit and with less or no affinity to re-

ceptors containing α2, α3 or α5 subunits (Pritchett & Seeburg, 1990). Crestani et al. (2000)

have demonstrated that the anticonvulsant and sedative effects mediated by zolpidem are

solely conferred by the α1 subunit. The described compounds are subunit selective, howe-

ver, they cannot discriminate between different receptor subtypes. Therefore, native GABAA

receptor subtypes cannot be identified using these compounds.

The application of classical benzodiazepines is often accompanied by adverse side effects,

such as dependence, withdrawal or tolerance limiting their clinical suitability (Buffett-Jerrott

& Stewart, 2002; Möhler et al., 2002; Bateson, 2004; Agid et al., 2007). It was suggested

that this was due to the lack of selectivity for a given receptor subtype. Ligands acting

on smaller subpopulations of GABAA receptors, such as zolpidem, are thought to produce

specific actions and show fewer side effects compared with the classical benzodiazepine

drugs. Currently there are a number of subtype specific ligands under investigation, which

may substitute the use of classical benzodiazepines in the future to contain the occurrence

of side effects (see Möhler et al., 2002; Rudolph & Möhler, 2006; Whiting, 2006). These

are, for example, zaleplone, a compound with high affinity for α1 subunits, but decreased

sensitivity to α2, α3, or α5 subunits or dihydroquinoline that has shown to exhibit agonist

efficacy at α2 subunits, but lacks affinity for α1-containing receptors (Johnstone et al., 2004;

Rudolph & Möhler, 2006). L-655,708 is a memory enhancing drug preferentially binding to

the α5 subunit (Navarro et al., 2002; Rudolph & Möhler, 2006; Whiting, 2006).

GABAA receptors are also targeted by volatile and intravenous anaesthetics, which exhibit

three different modes of action on the receptors: (1) allosteric receptor activation at low

concentrations, (2) at higher concentrations they directly activate receptors in the absence of

11


GABA and (3) at very high concentrations they block the chloride channel (Thompson et al.,

1996). A number of studies have demonstrated the importance of β subunit isoforms for

the action of certain anaesthetics, such as etomidate (Harris et al., 1995; Sanna et al., 1995,

1997). Etomidate produces modulatory actions on recombinant receptors containing any of

the β subunits, but has been shown to be less potent on β1-containing receptors compared to

β2- and β3-containing receptors (Hill-Venning et al., 1997; Sanna et al., 1997). The direct

action of etomidate, in contrast, is dependent on the presence of β2 and/or β3 subunits, since

receptors containing the β1 subunit did not show GABA-mimetic actions (Belelli et al., 1996;

Hill-Venning et al., 1997; Pistis et al., 1997; Sanna et al., 1997). These findings have led to

the suggestion that the modulatory and GABA-mimetic actions of etomidate are mediated

via distinct binding sites on the GABAA receptor.

Similar studies as described above using genetically engineered mice have shown that the

different actions of anaesthetics are also mediated by different GABAA receptor subtypes.

Point mutations of residue N265 in β2 (N265S) and β3 (N265M) subunits were introduced

in mice. The effects of etomidate in these mice were compared to the effects seen in wild

type mice (Jurd et al., 2003; Reynolds et al., 2003). Missing or reduced behavioural pattern

in mutant mice was suggested to be mediated by the mutated subunit. Via this means it was

demonstrated that the β3 subunit mediates the immobilising as well as in part the hypno-

tic effect of etomidate. The β2 subunit was found to also mediate the hypnotic action of

etomidate (Figure 1.3).

Neurosteroids are the only endogenously synthesised GABAA receptor modulators. They

interact with many different receptor subtypes and lack specificity, although several reports

suggested a more potent action on extrasynaptic δ-containing receptors (Belelli et al., 2002;

Wohlfarth et al., 2002; Mitchell et al., 2008). It was proposed that neurosteroids mediate

their action via two different binding pockets (Belelli & Lambert, 2005; D’Hulst et al., 2009;

Olsen & Sieghart, 2009), since they enhance GABA-mediated effects at low concentrations

whereas at higher concentrations they directly activate GABAA receptors. The binding site

for the former action has recently been shown to consist of highly conserved amino acid

residues in the TM domains of α subunits, whereas the site for direct receptor activation is

12


located in the interface between α and β subunits (Hosie et al., 2007, 2009).

As described before, there are various further drug binding sites located at the GABAA re-

ceptor (for a detailed review see Korpi et al., 2002) and the examples described illustrate the

complexity of allosteric ligand binding on GABAA receptor subtypes. They also indicate

the great potential of GABAA receptors for future drug development, especially in terms of

subtype selective compounds. Currently, most drugs in use are not subtype selective, but act

on multiple subunits/receptor subtypes. Consequences are frequent appearance of adverse

side effects. Hence, new subtype-selective compounds are needed.

1.2.5 Minor GABAA receptors

As mentioned above, most GABAA receptors comprise of α, β and γ subunits. However, in

certain areas of the brain receptors containing minor subunits are expressed and may play

important roles. ε, θ and π subunits were the last identified GABAA receptor subunits. They

were discovered by searching DNA databases (Davies et al., 1997; Hedblom & Kirkness,

1997; Whiting et al., 1997; Bonnert et al., 1999). All three have been assigned into their

own subunit classes, due to the relative low sequence similarities with members of other

subunit classes. ε and θ subunits are closest to γ and β subunits (Section 1.2.1), whereas

the π subunit shows closest homology to β (37 % identity), δ (35 % identity) and ρ (33 %

identity) subunits (Hedblom & Kirkness, 1997).

Little is known about properties and functions of these subunits and their roles in vivo are

still not known. It seems as if there has been a lack of interest in investigating these su-

bunits, since only a relatively small number of studies have been performed, compared to

other GABAA receptor subtypes. This may be partly, due to the fact that receptor subtypes

containing ε, θ or π subunits are insensitive to benzodiazepine-type drugs and no subunit-

specific pharmacologies could yet be established. Furthermore, their expression patterns are

restricted to few brain areas or, in case of the π subunit, are mainly expressed in peripheral

tissues and barely in the brain (Hedblom & Kirkness, 1997). However, these subunits may

have important functions in some areas of the brain.

13


1.2.5.1 The GABAA receptor ε subunit

The ε subunit was first described in 1997 by Davies et al., Garret et al. and Whiting et al.. It

was detected in several peripheral tissues, such as heart, liver, testis and placenta, as well as

in the mammalian brain, where its expression pattern is restricted to specific areas (Whiting

et al., 1997; Erlitzki et al., 2000; Davies et al., 2002). In situ hybridisation and Northern

blot analyses in primates and rodents have revealed distinct ε subunit expression in several

areas of the forebrain, such as basal ganglia, thalamus, hypothalamus and amygdala as well

as in the locus coeruleus (LC; Whiting et al., 1997; Moragues et al., 2000; Sinkkonen et al.,

2000; Sieghart & Sperk, 2002). ε subunit gene expression has also been demonstrated in

all forebrain and a few hindbrain cholinergic cell groups; many of these also contain the α3

subunit (Moragues et al., 2002). Furthermore, the expression pattern of the ε subunit widely

overlaps with that of the θ subunit in the brain, for example in hypothalamus, amygdala and

LC (Bonnert et al., 1999; Sieghart & Sperk, 2002). The θ subunit also exhibits a restricted

expression pattern in the brain. It is noteworthy that both subunit genes as well as the one

encoding the α3 subunit are clustered together on chromosome Xq28 in the human genome

(see Section 1.2.2; Wilke et al., 1997). Therefore, it is likely that the three subunits alone or

containing a further subunit form functional receptor channels in some areas of the brain, for

example in the LC or other areas in the forebrain, where all three subunits are co-expressed

(Moragues et al., 2000; Sieghart & Sperk, 2002).

Kasparov et al. (2001) proposed that neurons of the caudal nucleus of the solitary tract ex-

press ε-containing receptors. The authors investigated rat brain slices electrophysiologi-

cally as well as performed immunocytochemistry analyses, and concluded that the obser-

ved benzodiazepine insensitivity of these neurons could be due to expression of ε-subunit-

containing receptors. Another area where ε-subunit-containing receptors are expressed is the

LC. GABAA receptors expressed in the LC have interesting properties. It was assumed that

they comprise of α3, θ, ε and probably β1/3 subunits, but do not contain γ subunits (Chen

et al., 1999; Belujon et al., 2009). Indeed, in situ hybridisation experiments in brains from

21-day old rats did not detect γ subunit expression in LC neurons. However, conflicting

data were obtained when investigating benzodiazepine sensitivity in these neurons. Chen

14


et al. (1999) reported diazepam did not effect the GABA-induced firing rate of LC neurons

using LC slice preparations. Belujon et al. (2009) performed in vivo extracellular and in vitro

patch-clamp recordings in adult rats and slice preparations from juvenile rats, respectively,

to investigate the effects of flunitrazepam and diazepam on LC neurons. The former had

no effect, but surprisingly, diazepam at micromolar concentrations was able to potentiate

GABA-induced currents in vitro. Functions of ε-subunit-containing receptors in these brain

regions are not known yet.

The ε subunit exhibits a great level of sequence divergence across species. Furthermore,

several ε subunit splice variants have been found (see also Section 3.1.2, Chapter 3; Whi-

ting et al., 1997). Northern blot analysis revealed an incomplete ε subunit splice variant of

5.5 kilobases (kb) in several peripheral tissues, such as placenta and pancreas, and a correctly

spliced messenger RNA (mRNA) of 3.5 kb in several brain areas, and, to a lesser extent, in

the heart. Davies et al. (2002) showed that only the 3.5 kb version of the ε subunit mRNA

results in a full length protein. Exclusively in rodents, a splice variant containing a large

Pro/Glx domain that consists of 483 amino acid residues (mostly proline/glutamic acid and

proline/glutamine tandem repeats) has also been recognised as well as a splice variant la-

cking this insertion (Moragues et al., 2000; Sinkkonen et al., 2000). Even without this large

insertion at the amino-terminus, the amino acid sequence homology between the human and

rodent ε subunits is only 68 %, whereas it is between 88 % and 99 % for other orthologous

GABAA receptor subunits from different species.

In recombinant expression systems, the ε subunit forms functional receptor channels when

co-expressed with α and β subunits (Davies et al., 1997; Whiting et al., 1997). It may, the-

refore, substitute for the γ or δ subunit in native receptors. However, two recent studies

have examined alternative stoichiometry in ε-subunit-containing receptors (Jones & Hender-

son, 2007; Bollan et al., 2008). Most studies expressing recombinant ε-containing GABAA

receptors used subunit ratios of 1:1:1 for α, β and ε subunits. Jones & Henderson (2007)

investigated the influence of a range of transfection ratios of ε versus α and β subunits and

discovered that spontaneous current density, receptor deactivation time and a shift in rectifi-

cation from outward to inward changed with increased ε subunit expression. It was suggested

15


that, unlike in recombinant αβγ receptors, which exhibit a fixed stoichiometry of two α, two

β and one γ subunit (Baumann et al., 2002), receptor stoichiometry may vary, depending on

the amount of ε subunit expression. It was proposed that higher ε subunit abundance may

create receptors containing two ε subunits, where one subunit substitutes for the γ subunit

and the second for a β subunit (Figure 1.4 C).

A different attempt to solve ε-containing receptor stoichiometry was approached by Bollan

et al. (2008), who used concatenated subunits. These subunit constructs, consisting of two

or three different subunits, enable investigation of receptor stoichiometry and characteristics

of receptors with different subunit isoforms in specific locations (Minier & Sigel, 2004).

Bollan et al. (2008) proposed, upon functional expression of different concatenated subunits

in Xenopus laevis oocytes, that the ε subunit can assemble at different positions in receptor

pentamers. However, most likely it is located at position four in α1β2ε receptors (α1-β2-

α1-ε-β2), substituting for the β2 subunit, or at position one in α1β2γ2ε receptors (ε-β2-α1-

β2-γ2), substituting for the α1 subunit (Figure 1.4 A, B).

α1α1α1α1

α1α1α1α1

β2β2β2β2

εεεε

β2β2β2β2

1

2

34

5εεεε

α1α1α1α1

β2β2β2β2

β2β2β2β2

γ2γ2γ2γ2

1

2

34

5α1α1α1α1

α1α1α1α1

β2β2β2β2

εεεε

εεεε

1

2

34

5

A B C

Figure 1.4: Schematic representation of possible subunit arrangement in ε-containingGABAA receptors. Subunit positions in (A) α1β2ε receptors, (B) α1β2γ2ε receptors (bothsuggested by Bollan et al., 2008) and (C) α1β2ε receptors (proposed by Jones & Henderson,2007). Yellow circles represent the agonist GABA.

As mentioned above, subunit position in the receptor pentamer contributes to different re-

ceptor properties, for example agonist binding. The proposed ε subunit positions in αβε

or αβγε receptors as well as in receptors containing two ε subunits would all result in re-

ceptors with no benzodiazepine binding site and only one GABA binding site (Figure 1.4).

The latter could explain the low Hill coefficient seen in ε-containing receptors exposed to

16


GABA (Whiting et al., 1997; Neelands et al., 1999; Jones et al., 2006). However, the subunit

combination and stoichiometry of ε-containing GABAA receptors in vivo is still unknown.

The ε subunit has often been described as γ-like, due to its highest sequence similarity with γ

subunits and the fact that it forms functional channels with α and β subunits in recombinant

expression systems (Whiting et al., 1997; Bollan et al., 2008). Jones & Henderson (2007)

demonstrated that the ε subunit exhibits β-subunit-like characteristics too, in particular β3-

subunit-like. A series of experiments showed that the ε subunit is able to access the cell

surface on its own as well as in conjunction with the α1 subunit, but not together with the

β2 or γ2 subunit. This is also the case for β subunits. Moreover, the ε subunit possesses

similar assembly domains as the β3 subunit (Jones & Henderson, 2007). This could enable

the ε subunit to also assemble at a different position in the receptor complex and to replace

a β subunit. However, unlike the β3 subunit, the ε subunit expressed alone or together with

the α1 subunit was unable to form functional, GABA-sensitive channels at the cell surface

(Jones & Henderson, 2007; Bollan et al., 2008).

ε-containing GABAA receptors possess unique pharmacological properties, for instance in-

sensitivity to benzodiazepines (mentioned above; Davies et al., 1997; Whiting et al., 1997).

This is not surprising, because benzodiazepine binding requires amino acid residues from

α and γ subunits (Section 1.2.4). Several studies have been conducted to acquire the phar-

macological profile of ε-containing GABAA receptors (Davies et al., 1997; Whiting et al.,

1997; Thompson et al., 1998; Neelands et al., 1999; Maksay et al., 2003; Ranna et al., 2006).

Results obtained examining the effect of general anaesthetic agents, such as propofol, pen-

tobarbital or etomidate on ε-containing receptors were conflicting. Investigations in na-

tive neurons expressing ε subunits have shown reduced sensitivity to anaesthetics (Irnaten

et al., 2002; Sergeeva et al., 2005). Davies et al. (1997, 2001) also reported that ε-subunit-

containing receptors are insensitive to pentobarbital and propofol in heterologous expression

systems. In contrast, several other studies showed that general anaesthetics potentiate the

effect of submaximal GABA concentrations (Whiting et al., 1997; Thompson et al., 1998;

Neelands et al., 1999). Since the level of enhancement was similar in αβε and αβγ receptors,

but lower compared to αβ receptors, it was suggested that the inclusion of a third subunit

17


into the receptor complex lowers the potency of anaesthetic agents, such as pentobarbital.

Thompson et al. (2002) investigated the discrepancies in sensitivity to general anaesthetics

between the different in vitro studies described above. It was concluded that they were cau-

sed by the use of different expression vectors, which resulted in an overexpression of the ε

subunit versus α and β subunits in the studies by Davies et al. (1997, 2001). Thompson et al.

(2002) presumed that receptor stoichiometry varied in the different studies, which was corro-

borated by a study mentioned earlier by Jones & Henderson (2007). Nevertheless, all studies

demonstrated direct receptor activation after applying high concentrations of the anaesthetic

agents.

Recombinant ε-containing receptors are highly sensitive to GABA, but they also exhibit

agonist-independent channel activity (Neelands et al., 1999; Davies et al., 2001; Maksay

et al., 2003; Wagner et al., 2005). Spontaneous currents have also been seen in homomeric

β1 receptors or heteromeric receptors containing the β1 and α2 or γ2 subunits, but not in

receptors containing all three subunits (Miko et al., 2004). Thus, αβε receptors are the first

receptors comprising of three different subunits that exhibit constitutive receptor activity in

vitro. Davies et al. (2001) reported that receptors containing α, β, ε and γ subunits also exhi-

bited spontaneous receptor activity. A possible reason for this may be that the open state of

ε-containing receptors is energetically more favourable than the closed receptor form (Nee-

lands et al., 1999; Maksay et al., 2003). Whether or not ε-containing receptors in vivo exhibit

spontaneous activity remains unclear. In vitro studies by Belujon et al. (2009), Irnaten et al.

(2002) and Sergeeva et al. (2005) in LC neurons, native parasympathetic cardiac neurons and

hypothalamic neurons, respectively, could not detect significant spontaneous receptor acti-

vity, whereas Jones et al. (2006) described the occurrence of constitutive receptor activity in

cultured ganodotrophin-releasing hormone neurons, which was comparable to that described

in human embryonic kidney (HEK) 293 cells expressing α2β3ε receptors.

From numerous studies investigating the pharmacological profile of ε-subunit-containing

receptors, it can be summarised that the receptors exhibit different pharmacological charac-

teristics compared to receptors expressing α and β subunits only or αβγ receptors (Davies

et al., 1997; Whiting et al., 1997; Thompson et al., 1998; Neelands et al., 1999; Maksay et al.,

18


2003; Ranna et al., 2006). However, to this date no ε-subunit-specific pharmacology could

be identified. Moreover, it is not known whether ε-subunit-containing receptors are located

at the synapse or extrasynaptic and subunit composition in ε-containing receptors in vivo as

well as the function of the ε subunit remain still elusive.

1.3 Investigating ion channel properties

Ion channels are involved in a wide range of different diseases, for example epilepsy. Over

recent years, pharmaceutical industries developed an enhanced interest in ion channel drug

targets, due to an increased availability of chemical compounds and the advancement in the

field of genomic research, providing a large amount of previously unknown drug targets.

Among the 100 top selling drugs worldwide are, at present, 15 % ion channel modulators

generating more than six billion dollars yearly (Gill et al., 2003). As mentioned above,

GABAA receptors exhibit a wide range of structural diversity and possess a number of dif-

ferent modulatory binding sites that are targeted by a plethora of different drugs. In order to

lessen the appearence of adverse side effects commonly implicated with current drugs on the

market, new subtype-selective drugs are needed. Many methods are available to study ion

channel/GABAA receptor properties, including expression studies, assays to gain structural

or pharmacological insights, gene targeting, for instance knockout mice, as well as electro-

physiological studies to directly assess ionic current flow (for a review see Smith & Simpson,

2003).

Although it is technically challenging to screen chloride channels for modulatory com-

pounds, over recent years, a number of assay formats have been developed to screen for

compounds that target ion channels/GABAA receptors. Important assay characteristics are

sensitivity, specificity, throughput and information content, among others (Xu et al., 2001).

Many of these thwart each other, as pictured in Figure 1.5. For instance, highly informa-

tive assays, such as electrophysiological methods, usually have a low throughput, whereas

high throughput techniques, for example fluorescence-based techniques, lack informational

content. Thus, it is important to select the right assay type for a specific project.

19


Figure 1.5: Interrelations between different attributes of ion channel assays. Solid lines markantagonising features; dashed lines display other relationships between them (from Xu et al.,2001).

There are biochemical as well as cell-based assays to examine ion channel properties or

screen for compounds. However, the latter are often the preferred method since they provide

a physiologically relevant environment (An & Tolliday, 2009). Several possibilities exist to

express functional channels for ion channel screening, such as primary cell culture or engi-

neered cell lines. Relevant and frequently applied biological systems for GABAA receptor

expression are, for example, mammalian cells or Xenopus laevis oocytes (Sections 1.3.1 and

1.3.2). As mentioned before, the choice of a suitable assay approach is a further important

factor when investigating ion channels. Electrophysiological methods, like the two-electrode

voltage-clamp and the patch-clamp technique (Section 1.3.3), measure ionic conductance ac-

cross the cell membrane and provide high quality and informational content when assessing

ion channel function and drug-channel interactions (Xu et al., 2001). However, only a limited

number of cells can be examined per experiment, because the method is technically challen-

ging and labour intensive yielding only 10-20 data points per day. It is, therefore, used in

basic research as well as for secondary or safety screenings applying a smaller number of

compounds (Gonzalez & Maher, 2002). However, since the method provides high quality

data, systems for automated electrophysiological recordings have been developed to improve

20


throughput. Nevertheless, further miniaturisation and, thus, improving throughput remains

difficult, due to the complexity of the assay setup (Xu et al., 2001). Miniaturisation plays an

important role in high throughput screening (HTS), since a smaller assay scale requires less

reagents, hence, reducing the overall costs. Automated recording methods are implicated

with high costs that limit their use in HTS.

Several alternative assay technologies with high throughput have been developed to identify

ion channel modulators from large chemical compound libraries. In most cases, ion flux

assays or fluorescence-based methods, such as chloride-sensitive indicator dyes, membrane

potential fluorescence probes or the use of halide-sensing mutant green fluorescent proteins

(GFP), are applied in preliminary, large-scale screening assays (Xu et al., 2001). Both can

be set up in microtiter plate format with the possibility of measuring multiple samples at the

same time (see Sections 1.3.4 and 1.3.5 and Sections 4.1.2 and 4.1.3, Chapter 4).

Many factors influence the choice of the assay system to investigate ion channel function and

to screen for compounds that modulate them, which include the level of sensitivity required,

assay resolution, safety issues, costs, the ability to handle low volumes and the availability

of a suitable measuring device.

1.3.1 Mammalian expression systems

Heterologous expression systems, such as mammalian cell lines or Xenopus laevis oocytes

(Section 1.3.2) are commonly used to investigate structure-function relationships of recom-

binant proteins (Sigel & Minier, 2005). Studying a protein in a foreign host cell rather than

in its genuine environment is advantageous because of the relative ’pure’ environment and

the possibility of choosing the cell type for expression according to the type of experiment

planned as well as practical and cost reasons.

Mammalian cell lines are a popular expression system for recombinant proteins. Plasmid

vectors carrying genetic information of the gene of interest can be introduced stably or tran-

siently into the cells which in turn translate, modify and express the desired protein (Witchel

21


et al., 2002). Producing stably transfected cell lines requires great initial effort and time. Ho-

wever, once produced they are easier and more convenient to use than transiently transfected

cells. Advantages of a stable cell line are the homogeneous cell population as well as the

fact that errors in protein assembly are reduced. However, if several different proteins are

being investigated, it would be necessary to prepare stable cell lines for each protein of inter-

est. In this case, it is often more practical to transiently transfect the different proteins into

the cell line of choice. Mammalian expression systems benefit from the fact that numerous

cells can be transfected at the same time. Also advantageous is that mammalian cell lines

may resemble the physiological system in humans more than alternative expression systems

and experiments can be carried out at physiological temperature (37°C; Witchel et al., 2002;

Hogg et al., 2008). However, experiments using transiently transfected cells often suffer

from a change in receptor expression over time as well as relatively low transfection effi-

ciencies. For certain applications, such as electrophysiological measurements, this problem

can be overcome by using a detection system for successfully transfected cells. For instance,

a plasmid that expresses GFP can be simultaneously transfected with the protein of interest,

highlighting transfected cells.

The HEK293 cell line is a frequently used expression system for recombinant proteins, such

as ligand-gated ion channels (Thomas & Smart, 2005). HEK293 cells exhibit similarities

with early differentiating neurons by expressing many mRNAs which are present in neurons,

such as neurofilament proteins. The cells also contain several signalling pathways found in

native neurons. Therefore, the expression of foreign neuronal proteins in HEK293 cells re-

sembles expression in their native cell environment (Thomas & Smart, 2005). Furthermore,

HEK293 cells are relatively easy and cheap to maintain, they are robust, have a quick cell di-

vision rate (between 24 to 36 hours) and a good size for electrophysiological measurements.

Several different transfection techniques have been applied successfully, resulting in suffi-

cient protein expression for many different applications, ranging from electrophysiological

recordings to biochemical assays. HEK293 cells express a number of endogenous receptors

(for details see Thomas & Smart, 2005), and experiments should be carefully monitored for

interference of these proteins using adequate controls. Nevertheless, the expression levels

22


of recombinant proteins are usually high enough that endogenous proteins are outnumbered

and the risk of interference is minimal.

1.3.2 Xenopus laevis oocytes

The oocytes of the South African clawed frog Xenopus laevis have been commonly used as

a heterologous expression system for over 30 years (Sigel & Minier, 2005), for example for

comparing the functions of wild-type and mutated ion channels or to investigate drug actions

on expressed proteins. Oocytes undergo six developmental stages (I-VI), which take place

asynchronous, i.e. the ovary contains oocytes of all stages (Dumont, 1972). Most electro-

physiological studies are performed using oocytes from stages V and VI. Characteristic for

Xenopus oocytes is the ’two-toned colour scheme’: the animal pole containing the nucleus

is dark brown whereas the vegetal pole exhibits a yellowish colouration (Figure 1.6).

Figure 1.6: Xenopus laevis oocytes (from Sigel & Minier, 2005).

Oocytes are able to express foreign genetic material precisely and at high levels (Sigel, 1990;

Sigel & Minier, 2005). Complementary DNA (cDNA) can be microinjected into the nucleus

of an oocyte or, alternatively, RNA can be introduced into the cytoplasm (Figure 1.7). After-

wards, the nucleic acid is efficiently transcribed (in the case of cDNA injection) and transla-

23


cDNA cDNA

RNA mRNA

protein

transcription

translation

microinjection

microinjection

nucleus

Figure 1.7: Schematic representation of the functional expression of genetic information inXenopus oocytes (after Sigel, 1990).

ted by the oocyte. The resulting protein is then assembled, post-translationally modified and

finally targeted to the surface membrane or an appropriate subcellular compartment.

Proteins are usually expressed within 24 hours after RNA or cDNA microinjection and oo-

cytes can then be used for up to 14 days (Wagner et al., 2000; Sigel & Minier, 2005). RNA

injection usually results in protein expression in > 98 % of the injected cells, whereas mi-

croinjection of cDNA is often successful in only a small percentage of the injected oocytes

(Sigel & Minier, 2005). There are several advantages of using oocytes over mammalian cells.

Oocytes are available at low cost and easy to handle in the laboratory due to their large size

(1-1.2 mm in diameter) and robustness (Wagner et al., 2000; Sigel & Minier, 2005). Further-

more, the number of endogenously expressed (surface) proteins is relatively low (Dascal,

1987). The quantity of electrophysiological data which can be obtained from a single oocyte

is relatively high; the robust setup permits recording for more than an hour (Witchel et al.,

2002). However, there are a number of limitations to the use of oocytes: (1) each oocyte

needs to be injected separately, (2) oocytes are less suitable for studying very fast cellular

events, due to their large size and associated slow fluid exchange rate and slow changes in

membrane potential, (3) the intracellular ionic composition cannot be influenced, (4) com-

pounds can be absorbed by the egg yolk, which may result in lower sensitivity, (5) oocytes

show seasonal differences in expression and quality and (6) frogs are poikilothermic ani-

24


mals. For that reason, experiments need to be carried out at room temperature rather than

at physiological temperature of 37°C (Wagner et al., 2000; Xu et al., 2001; Sigel & Minier,

2005; Hogg et al., 2008). This may hamper certain applications, for instance kinetic studies.

Moreover, findings should be evaluated carefully, since there may be variations between the

expression in oocytes and the native tissue, as a result of different signalling pathways.

1.3.3 Electrophysiological techniques

Electrophysiological techniques measure differences in current carried by ions accross a cell

membrane of a whole cell or a membrane patch (Xu et al., 2001). The two-electrode voltage-

clamp technique is often used to study ion channel properties in large cells, such as Xenopus

laevis oocytes. The technique measures whole-cell currents by impaling two glass microelec-

trodes into an oocyte; one electrode acts as a current-delivering electrode (i), which clamps

the membrane potential to a desired voltage, and the other measures the membrane poten-

tial (V; see Figure 1.8). The measured membrane potential is compared with the command

voltage (delivered by the current delivering electrode) and the difference is set to zero by the

amplifier. Thus, it is possible to measure ionic current flow across the membrane (deviation

from the baseline current) mediated by either ion channels or transporters (see also Section

1.3.2; Baumgartner et al., 1999; Wagner et al., 2000).

The patch-clamp method is the ’gold standard’ for measuring ion channel functions, pro-

viding high quality and physiologically relevant data. It is a highly flexible technique with

different configurations, such as whole-cell recording, inside-out and outside-out patches

(Hamill et al., 1981). The first configuration allows the recording of ionic currents from

ion channels expressed in the whole cell membrane (Figure 1.9), whereas the latter two are

single channel configurations. They enable the study of individual ion channels in isolation

in a small membrane patch to obtain, for example, data on channel gating and kinetics. The

main difference to the two-electrode voltage-clamp technique is that a single micropipette

is pressed against the membrane of a cell forming an electrical "giga-seal" (high resistance

seal in the gigaohm range) between the membrane and the micropipette. Giga-seals enable

most of the ionic currents to go into the pipette, reducing the ’leak current’; they are me-

25


Figure 1.8: Schematic representation of the two-electrode voltage-clamp technique.The membrane potential is measured by an electrode (V) that is linked to a voltage follower (triangle)and it is then compared with the command voltage (E; given by a generator) via the feedback amplifier(FBA). The amplified difference between the signals is delivered as a current through the current-delivering electrode (i), which clamps the membrane potential to a desired voltage and further overthe cell membrane to the grounding electrode in the bath. Changes in the ion flux across the membranecan then be monitored as variation from the baseline current (adapted from Wagner et al., 2000).

26


chanically stable and offer a low background noise (Hamill et al., 1981). The advantages

of this technique are that the intracellular environment can be controlled, due to the mixing

of pipette solution and cytosol (whole-cell recording), it can be performed at physiological

temperatures and fast cellular events can be studied, due to a rapid solution exchange rate

and a high temporal resolution.

Cell

Patch pipette

Pipette solution

Figure 1.9: Schematic representation of the patch-clamp whole-cell configuration.This method enables the recording of ionic currents from ion channels expressed in the whole mem-brane of the cell. To acquire the whole-cell configuration, a glass microelectrode is lowered ontothe cell surface (cell-attached mode), slight suction is then applied to obtain a high-resistance sealbetween cell membrane and glass pipette. Subsequently, the membrane is ruptured and the amplifiercan control and measure voltage accross the entire cell membrane (adapted from Wood et al., 2004).

1.3.4 Flux assays

Commonly used ion flux assays apply radioactive tracer ions, such as 36Cl- or 125I- which can

be used to measure ion influx or efflux through ion channels, thereby measuring ion channel

activity directly (Smith & Simpson, 2003; Terstappen, 2005; Verkman & Galietta, 2009).

The throughput of these assays is considerably higher than with electrophysiological recor-

dings, since assays are usually performed in 96- or 384-well plates, however, it is lower com-

pared to fluorescence-based methods. Radioactive flux assays are easy to perform, reliable

27


and robust and exhibit high sensitivity. Nevertheless, slow temporal resolution, reduced re-

producibility, difficulties of further assay miniaturisation and low informational content as

well as high costs and safety issues are major drawbacks (Gill et al., 2003). Tang & Wil-

dey (2004) recently presented a non-radioactive flux assay using iodide as an indicator of

chloride channel activity (Section 4.1.2, Chapter 4). The advantages of this assay setup are

lower costs and increased safety. However, sample mixtures and buffers of the assay are re-

latively unstable and fast reading is required, which may limit its use in HTS. Furthermore,

flux assays suffer from long incubation times needed to load the cells with tracer molecules

and require high levels of ion channel expression to lower the signal-to-noise ratio (Xu et al.,

2001).

1.3.5 Fluorescence-based assays

Fluorescence-based methods are ideally suited for assay miniaturisation, because they are

volume-independent. There are several different fluorescence-based methods to screen for

ion channel modulators, for instance chloride indicator dyes or membrane potential-sensing

dyes (Gill et al., 2003; Molokanova & Savchenko, 2008; Verkman & Galietta, 2009). Assays

applying membrane permeable, quinoline-derived chloride indicator dyes, such as MQAE,

are easy to set up and exhibit relatively high chloride sensitivity (Marandi et al., 2002).

However, they suffer from high background noise, leakage from the cells and the low half-

life of the fluorescence dye (Gill et al., 2003; Verkman & Galietta, 2009). Another method

is to measure changes in intracellular pH, using proton-sensitive dyes and a fluorometric

imaging plate reader (FLIPR; Gill et al., 2003; Verkman & Galietta, 2009). Voltage-sensing

Förster (or Fluorescence) Resonance Energy Transfer (FRET)-based, ratiometric dyes are

also commonly used for ion channel screening. FRET is the transfer of energy from an

excited fluorophore (donor) to another fluorescent molecule (acceptor), which is dependent

on the distance between them (Piston & Kremers, 2007). These dyes are lipophilic molecules

and can be easily introduced into the cell membrane. Changes in membrane potential can

then be measured with fast resolution using a voltage/ion probe reader; it is also possible to

quantify membrane potential changes. A wide range of different dyes is available to measure

28


changes in membrane potential. They differ in chloride sensitivity, response kinetics and

sensing mechanisms (Verkman & Galietta, 2009). The main drawback of these dyes is that

the membrane potential can be influenced by many different factors, such as cytoplasmic

composition or activity of other ion channels or transporters in the membrane rather than

only the channel of interest. The latter leads to a high rate of false positive readings which is

further increased by test compounds exhibiting autofluorescence. Furthermore, such systems

are more suitable for stably transfected cell lines, since the dye molecules are present in every

cell’s membrane. False negative results can also be obtained due to the high background

noise masking weaker signals.

There are also assay technologies using halide-sensing mutant GFPs, such as Clomeleon.

Clomeleon is a ratiometric, FRET-based and genetically-encoded chloride-ion indicator pro-

tein. It was developed by Kuner & Augustine (2000) and consists of a chloride-sensitive

yellow fluorescence protein (YFP) linked to a chloride-insensitive cyan fluorescence protein

(CFP) and can be targeted to specific cellular compartments (see Section 4.1.1, Chapter 4).

Compared to chloride-sensitive indicator dyes, such as MQAE, Clomeleon exhibits slower

kinetics and its chloride sensitivity is relatively low. YFP mutants with higher chloride sen-

sitivity have, therefore, been developed, such as YFP-H148Q/I152L (Section 4.1.3, Chapter

4; Galietta et al., 2001a; Verkman & Galietta, 2009). Although fluorescence membrane

potential-sensing dyes and YFP mutant proteins both have similar signal-to-noise ratios and

dynamic ranges, there are several advantages of YFP-based assays: they are faster, less ex-

pensive, YFP is genetically encoded and cannot leak out of the cells and the assay does

not require ion channel expression in a stable cell line, but can also be used in transiently

transfected cells, which is advantageous when investigating several of the various GABAA

receptor subtypes (Kruger et al., 2005).

1.4 Aims and objectives of the study

As mentioned in Section 1.2.5.1, expression of the GABAA receptor ε subunit is relatively

restricted in the mammalian brain; it is mainly expressed in the forebrain as well as in all

29


major neuronal groups, including cholinergic cells (Moragues et al., 2000, 2002). 84-95 % of

cholinergic neurons also express the α3 subunit (Gao et al., 1995). Due to the co-expression

of α3 and ε subunits in cholinergic neurons, it was hypothesised that the subunits form

functional channels in these neurons, which may regulate ACh release (McGehee & Role,

1996; Cervetto & Taccola, 2008). The α3- and ε-subunit-containing receptor subtype is,

therefore, a potential target for the development of anti-AD drugs that would function by

increasing ACh release in forebrain regions where cholinergic neurons are dying.

The initial aim of this project was to functionally express and pharmacologically characterise

the GABAA receptor subtype that contains the α3 and ε subunits applying an electrophysiolo-

gical approach. It was shown in several studies, that the β subunit has no significant effect on

the benzodiazepine modulation of the GABAA receptor (Pritchett et al., 1989b; Hadingham

et al., 1993). Therefore, the α3 and ε subunits were co-expressed with the rat β2 subunit, a

subunit combination that has not been pharmacologically characterised before. However, it

is likely that receptor subtypes of this combination exist in the brain, since the β2 subunit

is the most widely expressed β subunit in the brain. At first, it was planned to establish the

patch-clamp method at Nottingham Trent University using HEK293 cells transiently trans-

fected with rat cDNAs encoding the respective GABAA receptor subunits. This would have

also provided information about correct assembly of the subunits in HEK293 cells, impor-

tant for the subsequent cell-based screening assay setup (see below). However, problems

with the setup of the electrophysiological rig at Nottingham Trent University caused a slight

change in the project and electrophysiological recordings were performed in Xenopus laevis

oocytes microinjected with complementary RNA (cRNA) in the laboratory of Dr Ian Mellor

at the University of Nottingham instead.

After the pharmacological characterisation of the α3β2ε receptor, the aim was to establish

a cell-based screening system using HEK293 cells and search for compounds that modulate

the activity of this GABAA receptor subtype. It is generally accepted that modulation of the

agonist binding site of ligand-gated ion channels is not a productive avenue for drug disco-

very. Therefore, as part of the project, search was narrowed down to compounds that bind

elsewhere on this multi-subunit receptor complex. Whiting (1999) suggested the existence

30


of a new modulatory site on ε-containing receptors, since steroids, barbiturates and anaes-

thetics have no significant influence on these receptors. Since GABAA receptor subunits

exhibit strong sequence similarity to one another (Section 1.2.2), it was hoped to identify

a site between the α3 and ε subunits that is analogous to the benzodiazepine binding site

that is found at the interface between α and γ subunits. For this purpose, a search for lead

compounds from a chemical compound library that, in the presence of GABA, modulate

receptor activity was conducted. Such compounds could not only serve as potential future

drugs on ε-containing receptor subtypes, but may also help providing important information

to uncover the role of the ε subunit in future studies.

31

Chapter 2

Materials and Methods

2.1 General molecular biology techniques

2.1.1 Preparation of competent Escherichia coli (E. coli) cells

100 µl already competent E. coli XL1 blue cells were added to 5 ml lysogeny broth (LB) me-

dium (10 g/l peptone from casein, 5 g/l yeast extract, 10 g/l sodium chloride, Miller), supple-

mented with tetracycline (25 µg/ml) and incubated overnight in an orbital shaking incubator

(S150, Stuart) at 37°C and 200 rpm. On the next day, 2 ml of the overnight culture was ad-

ded to 100 ml LB medium supplemented with tetracycline (25 µg/ml) and incubated further

at 37°C and 200 rpm for several hours, measuring the optical density (OD) of the culture

at regular intervals at a wavelength of 660 nm until an OD660 of 0.4 to 0.5 was reached.

The inoculated LB medium was then split and transferred to two 50 ml tubes (Sartstedt) and

centrifuged 3,000 rpm for 15 minutes (Beckman Coulter) at room temperature (18-25°C).

Supernatants were discarded subsequently and the bacterial cell pellets resuspended in 20 ml

cold, sterile-filtered TFB1 buffer (30 mM KAc, 50 mM MnCl2, 100 mM KCl, 10 mM CaCl2

and 15 % (v/v) glycerol; pH 5.8) each by briefly vortexing the tubes. Next, the bacteria were

incubated on ice for 30 minutes, followed by centrifugation at 1,750 rpm for 20 minutes at

4°C. Supernatants were discarded and bacterial cell pellets carefully resuspended in 2 ml

ice-cold TFB2 buffer (10 mM MOPS buffer, 76 mM CaCl2, 10 mM KCl and 15 % (v/v)

glycerol). Finally, the competent E. coli Xl1 blue cells were aliquotted into microfuge tubes

32

Chapter 2. Materials and Methods

and stored at -80°C until usage.

2.1.2 Transformation of plasmid DNA into competent E. coli cells

1 to 5 µg plasmid DNA was added to 50 µl competent E. coli Xl1 blue cells and incubated

on ice for 30 minutes. The cells were then heat-shocked by incubating them at 42°C for

50 seconds and subsequently cooled on ice for another two minutes. 100 µl SOC medium

(20 g/l 2 % (w/v) Bacto tryptone, 0.5 % (w/v) Bacto yeast extract, 10 mM NaCl, 10 mM KCl,

20 mM MgCl2, 20 mM glucose) was added followed by incubation at 37°C, 200 rpm for 45

minutes. Transformed bacteria were cultured on LB agar plates (10 g/l peptone from casein,

5 g/l yeast extract, 10 g/l NaCl, 12 g/l agar-agar, Miller) containing 0.1 % (w/v) ampicillin

overnight at 37°C.

2.1.3 Plasmid preparation

2.1.3.1 Minipreparation

Plasmid DNA transformed into E. coli cells was isolated by selecting single colonies from

a LB agar plate using a sterile pipette tip. 2 ml LB medium supplemented with 0.1 % (w/v)

ampicillin were inoculated with a single bacterial colony and incubated overnight at 37°C,

200 rpm. On the following day, the plasmid DNA was purified from the overnight cultures

using a GenEluteTM Plasmid Miniprep Kit (Sigma) or QIAprep Spin Miniprep Kit (Qiagen)

following the manufacturer’s protocols. Isolated plasmid DNA was finally eluted in 50 µl

0.1 % (v/v) diethylpyrocarbonate (DEPC)-treated water and stored at -20°C.

2.1.3.2 Maxipreparation

In order to generate multiple copies of plasmid DNA, single bacterial colonies containing

the desired plasmid were transferred into 5 ml LB medium supplemented with 0.1 % (w/v)

ampicillin and incubated for 8 hours at 37°C, 200 rpm. Next, 2 to 3 ml of the bacterial culture

was transferred to 100 ml LB medium (containing 0.1 % (w/v) ampicillin) and incubated

overnight at 37°C, 200 rpm. On the following day the plasmid DNA was isolated using a

33


Plasmid Maxi Kit (Qiagen) following the manufacturer’s instructions. The purified DNA

was resuspended in 200 µl 0.1 % (v/v) DEPC-treated water and stored at -20°C for further

use.

2.1.4 Restriction endonuclease digestion

Purified plasmid DNAs and polymerase chain reaction (PCR) products were digested using

one or two (depending on further applications) restriction enzymes (Promega). First, enough

0.1 % (v/v) DEPC-treated water was added into 0.5 ml microcentrifuge tubes to make up a

final volume of 10 or 20 µl. Reactions were then further assembled by adding the respective

10 x buffer, 0.2 to 1.5 µg DNA and respective restriction endonuclease, using a tenfold ex-

cess of enzyme over DNA (see Section 2.2.3 for details). After incubating the mixtures for

1 to 2 hours at 37°C, the digested as well as corresponding undigested plasmid DNAs were

analysed by gel electrophoresis on a 1 % (w/v) agarose gel, containing ethidium bromide

(10 mg/ml), in 1× tris-acetate (TAE)-buffer (40 mM Trizma Base, 1.14 % (v/v) Glacial ace-

tic acid, 1 mM disodium EDTA (pH 8)) and visualised by UV trans illumination (InGenius

or G:Box, Syngene).

2.1.5 Measuring nucleic acid concentrations

Nucleic acids were diluted and concentrations were measured using a UV spectrophotometer

(DU530 or DU800, Beckman). Based on the absorbance (A) at 260 nm the concentrations

were calculated as shown below:

concentration [ng/µl] = A260 × dilution factor× conversion factor (2.1)

The conversion factors for nucleic acids are as follows (Maniatis et al., 1982):

Double-stranded DNA : 50 ng/µl

Single-stranded DNA & RNA : 40 ng/µl

Single-stranded oligonucleotides : 20 ng/µl

34


2.1.6 Ethanol precipitation

Nucleic acids were precipitated using 2.5 × volume 100 % (v/v) ethanol and 1/10 volume

3 M NaAc (pH 5.2), incubated at -20°C for at least 30 minutes and subsequently centrifuged

at 13,000 rpm, 4°C for 30 minutes. Supernatants were discarded and pellets were washed

with 75 % (v/v) ethanol and centrifuged again at 13,000 rpm, 4°C for five minutes. Nucleic

acid pellets were briefly air-dried before resuspension in DEPC-treated water.

2.1.7 Sequencing

15-20 µl DNA at a concentration of 60-100 ng/µl were added into a 1.5 ml microfuge tube.

Tubes containing DNA samples were sent to Eurofins MWG Operon (London) for sequen-

cing. Sequencing results were aligned and compared with corresponding GABAA receptor

subunit or other respective sequences in the nucleotide sequence database using a basic lo-

cal alignment search tool (BLAST; URL: http://www.ncbi.nlm.nih.gov/BLAST/) and a se-

quence alignment tool (ClustalW; URL: http://www.justbio.com/aligner/index.php), respec-

tively, to confirm their identity.

2.1.8 Side-directed mutagenesis - Generating a truncated ε subunit

2.1.8.1 Primer design

A truncated ε (εT) subunit version, lacking 26 amino acids ranging from within TM2 until

TM3 compared to the full length ε subunit version (Figure 3.1, Chapter 3), was construc-

ted using overlap PCR and cDNA encoding the rat ε subunit (in pcDNA3 vector). Two

sets of sequence specific primers were designed according to the ε subunit cDNA sequence

(Table 2.1).

Table 2.1: Oligonucleotide design to generate a GABAA receptor εT subunit

Oligonucleotide Sequence Tm

rGABRE-T_F1 5’-GCC AGT GTG CTG GAA TTC TG-3’ 56.6°CrGABRE-T_R1 5’-AGA AGT CCA AAG CTA CAG AGG CCC T-3’ 61.6°C

rGABRE-T_F2 5’-AGG GCC TCT GTA GCT TTG GAC TTC T-3’ 61.6°CrGABRE-T_R2 5’-CTA GAT GCA TGC TCG AGC GG-3’ 58.1°C

35


Forward and reverse primers (rGABRE-T_F1 and rGABRE-T_R2), flanking the entire ε

subunit sequence, overlapped with the multiple cloning site of the ε-subunit-containing vec-

tor and contained an EcoRI or XhoI endonuclease restriction site (underlined), respectively.

The introduction of restriction sites facilitated the insertion of the εT subunit cDNA into the

pcDNA3.1+ vector after generating the correct sequence. The oligonucleotides positioned

within the ε subunit sequence, rGABRE-T_R1 and rGABRE-T_F2, were flanking both sites

of the mutation which was located in the middle of the primers. All four oligonucleotides had

closely matched melting temperatures (Tm) for optimal amplification which were calculated

using the following website: URL:http://eu.idtdna.com/analyzer/applications/oligoanalyzer/

default.aspx. Figure 2.1 illustrates the different amplification steps used in order to generate

the respective deletion.

ATGdeletion

EcoRI

5’ 3’

PCR 1

ε subunit sequence in pcDNA3

PCR 3

EcoRI XhoIXhoI

εT

rGABRE-T_F1

rGABRE-T_R1

rGABRE-T_F1 rGABRE-T_R2

Product 2

XhoIXhoIPCR 2

rGABRE-T_R2

rGABRE-T_F2

Product 15’ 3’

3’5’

Figure 2.1: Side-directed mutagenesis using overlap PCR to create the εT subunit.The forward primer rGABRE-T_F1 is located before the start codon of the ε subunit sequence andcontains an EcoRI restriction site. The reverse primer rGABRE-T_R1 is spanning the sequence im-mediate before and after the 26 amino acid mutation, excluding the sequence which was deleted.The second primer pair consists of the reverse complement of rGABRE-T_R1 (rGABRE-T_F2) andrGABRE-T_R2. The latter is located after the stop codon of the ε subunit sequence extending into themultiple cloning site of the vector and including a XhoI restriction site. PCR products 1 and 2 wereused as templates in PCR 3, using oligonucleotides rGABRE-T_F1 and rGABRE-T_R2. rGABRE-T_R1 and rGABRE-T_F2 which are comprised in PCR products 1 and 2 anneal together resulting inthe amplification of the whole length εT subunit sequence.

36


2.1.8.2 Polymerase chain reaction

All PCRs were performed using Pfu DNA polymerase (Promega) or PhusionTM high fidelity

DNA polymerase (Finnzymes). Both enzymes exhibit proofreading activity to minimise er-

rors due to incorporation of wrong bases by the enzyme. PCR 1 and 2 (see Figure 2.1) were

assembled in sterile 0.5 ml PCR tubes as follows:

Pfu DNA polymerase 10 × buffer (200 mM Tris-HCl (pH 8.8 at 25°C),

100 mM KCl, 100 mM (NH4)2SO4, 20 mM MgSO4, 1.0% Triton® X-

100 and 1 mg/ml nuclease-free BSA, Promega)

5 µl

dNTP mix (10 mM each, Sigma) 1 µl

Forward primer (200 ng/µl) 1 µl

Reverse primer (200 ng/µl) 1 µl

DNA template (1 pg/µl (PCR 1) or 5 pg/µl (PCR 2)) 1 µl

pfu DNA polymerase (2-3 U/µl, Promega) 1 µl

DEPC-treated water to a final volume of 50 µl

Negative controls contained water instead of the DNA template. PCRs were performed in a

thermal cycler (TC-512 or TC-3000, Techne) at the following conditions:

Initial Denaturation 94°C 5 min

32 cycles

Denaturation 94°C 1 min

Annealing 59°C 1 min

Extension 72°C 4.5 min

Final extension 72°C 10 min

PCR products were visualised on a 1% agarose gel (Section 2.1.4) and amplicons of the right

size, ~960 bp (PCR 1) and ~2100 bp (PCR 2), were excised from the gel with a clean scal-

pel and purified using a GFXTM PCR DNA and Gel Band Purification Kit (GE Healthcare)

according to the manufacturer’s instructions. Purified PCR products were ethanol precipita-

37


ted (Section 2.1.6). The two PCR products then served as template DNA for a further PCR

(PCR 3) using the forward primer of the first (rGABRE-T_F1) and the reverse primer of the

second primer pair (rGABRE-T_R2) and PhusionTM high fidelity DNA polymerase (Finn-

zymes). The PCR was assembled as follows:

5 × Phusion® HF buffer (containing 7.5 mM MgCl2) 10 µl

10 mM dNTPs (200 µM each, Sigma) 1 µl

Forward primer (200 ng/µl) 1 µl

Reverse primer (200 ng/µl) 1 µl

purified PCR product 1 (1 in 10 dilution) 1 µl

purified PCR product 2 (undiluted) 0.5 µl

DMSO (100 %) 1 µl

Phusion® DNA Polymerase (2 U/µl) 0.5 µl


Negative controls contained water instead of purified PCR products. The PCR was carried

out in a Techne TC-3000 thermal cycler at the following conditions:

Initial Denaturation 98°C 30 sec

35 cycles

Denaturation 98°C 10 sec

Annealing 64°C 30 sec

Extension 72°C 1.5 min

Final extension 72°C 10 min

The resulting product was visualised on a 1% agarose gel. The amplicon of the right size

(~3000 bp) was excised from the gel, purified and ethanol precipitated (see above).

38


2.1.8.3 Ligation of the εT subunit cDNA into the pcDNA3+ vector

The purified εT subunit cDNA (PCR product 3) was digested in a 0.5 ml microfuge tube as

follows:

Purified PCR product 3 14.5 µl

pcDNA3.1+ vector (20 ng/µl) 1.5 µl

10 × buffer H (0.9 M Tris-HCl, 100 mM MgCl2, 0.5 M NaCl, pH 7.5, Promega) 2 µl

EcoRI (12 U/µl, Promega) 1 µl

XhoI (10 U/µl, Promega) 1 µl


The sample was incubated for 2 hours at 37°C before it was again ethanol precipitated to

purify the DNA (Section 2.1.6). Afterwards, the digested PCR product was ligated into the

linearised pcDNA3.1+ vector in a final volume of 10 µl, by adding 1 µl T4 DNA ligase (1-

3 U/µl, Promega) and 1 µl 10× reaction buffer (300 mM Tris-HCl (pH 7.8 at 25°C), 100 mM

MgCl2, 100 mM DTT and 10 mM ATP, Promega) and incubating it at 4°C overnight. On

the following day, the ligated product (10 µl) was transformed into 50 µl competent E. coli

cells (Section 2.1.2). Bacterial colonies, containing plasmid DNA, were cultured overnight

and subsequently purified using GenEluteTM Plasmid Miniprep Kit (Section 2.1.3). 0.5 µl

of purified plasmid DNA was digested using EcoRI and XhoI in a final volume of 20 µl as

described above and subsequently visualised on a 1% agarose gel in order to display the

insertion of the εT subunit DNA into the vector. Samples, which contained an insert, were

sent for DNA sequencing (Section 2.1.7) to confirm their correct identity and orientation in

the pcDNA3.1+ vector.

39


2.2 GABAA receptor subunits & fluorescence proteins

2.2.1 Recovery of vectors harbouring GABAA receptor subunits

Rat GABAA receptor α3, β2, γ2L and ε subunit cDNAs were kindly provided by Prof. Erwin

Sigel (Institute of Biochemistry and Molecular Medicine, University of Bern, Switzerland)

and Dr Maurice Garret (Laboratoire de Neurophysiologie, Bordeaux, France). The α3 subu-

nit cDNA was subcloned into the pcDNA3.1+ vector (Invitrogen), the β2 and γ2L subunit

cDNAs were in the pCMV vector and the ε subunit cDNA was present in the pcDNA3 vector

(Invitrogen; see Appendix A for vector maps). Plasmids that harboured the α3, β2 and γ2L

cDNAs were recovered from filter paper by cutting out the area on the paper, where the DNA

was placed, with sterile scissors and transferring the paper to sterile 1.5 ml microfuge tubes.

Next, 50 µl 10 mM Tris-HCl (pH 7.6) was added to the tubes; they were vortexed, incubated

at room temperature (18-25°C) for 5 minutes and briefly centrifuged for 30 seconds. The

supernatants containing the plasmid DNAs were transferred into fresh tubes and stored at

-20°C for further use. The ε-subunit-containing vector was obtained as aqueous stock solu-

tion (0.2 mg/ml) and stored at -20°C. The human GABAA receptor α1 subunit was present

in the pBS vector (Stratagene) and stored in ethanol at -20°C.

2.2.2 Recovery of pRK5-Clomeleon and pEYFP H148Q/I152L vectors

The pRK5-Clomeleon plasmid was generously provided by Dr Thomas Kuner (Institute of

Anatomy and Cell Biology, Heidelberg University, Germany) and recovered from filter paper

as described above (Section 2.2.1), but instead of Tris-HCl, 50 µl 0.1 % (v/v) DEPC-treated

water was added to the filter paper in the microfuge tube. This was vortexed, incubated

at room temperature (18-25°C) for 5 minutes and briefly centrifuged for 30 seconds. The

supernatant, containing the plasmid DNA, was transferred to a fresh tube and stored at -

20°C until further use.

A pcDNA3.1+ vector containing the coding sequence for the pEYFP H148Q/I152L mutant

protein was kindly supplied by Prof. Alan Verkman (Department of Medicine, University

of California, San Francisco, USA). The vector was obtained as aqueous stock solutions

40


(0.2 mg/ml) and stored at -20°C.

2.2.3 Examination of received vectors

Before the cDNAs described above (Sections 2.2.1 & 2.2.2) could be used in experiments,

their sequence identity was analysed. On this account, the plasmid vectors were transformed

into competent E. coli Xl1 blue cells and subsequently purified and, in case of the GABAA

receptor subunits, digested using restriction endonucleases (Promega) listed in Table 2.2

(Sections 2.1.2 to 2.1.4). After digesting the plasmid cDNAs, they were visualised on a 1%

agarose gel (Section 2.1.4) to verify the presence of GABAA receptor subunit cDNA in the

vector. Finally, GABAA receptor subunit as well as pEYFP H148Q/I152L cDNAs and the

pRK5-Clomeleon plasmid were sent for sequencing (Section 2.1.7).

Table 2.2: Restriction enzymes used to digest different GABAA receptor cDNAs

Enzymes 10 × buffer (buffer composition)α3 HindIII, XbaI Buffer B (60 mM Tris-HCl, 60 mM MgCl2, 0.5 M NaCl, 10 mM DTT, pH 7.5)β2 EcoRI, XbaI Buffer H (0.9 M Tris-HCl, 100 mM MgCl2, 0.5 M NaCl, pH 7.5)γ2L EcoRI, XbaI Buffer Hε EcoRI, XbaI Buffer Hα1 HindIII, EcoRI Buffer E (60 mM Tris-HCl, 60 mM MgCl2, 1 M NaCl, 10 mM DTT, pH 7.5)

2.2.4 Subcloning

The sequencing results revealed that the α3 subunit cDNA was present in the wrong orienta-

tion in the pcDNA3.1+ vector and had to be reversed. The α1 subunit cDNA was present in

the pBS vector and needed to be re-cloned into a mammalian expression vector (pcDNA3.1+)

in order to express the subunit in mammalian cells in later experiments. Furthermore, for the

purpose of cRNA synthesis using T7 RNA polymerase (Section 2.3.1), the β2 and γ2L su-

bunit cDNAs were sub-cloned into the pcDNA3.1+ vector. Plasmids containing the GABAA

receptor subunit cDNAs were digested using restriction endonucleases as described before

(Sections 2.1.4 and 2.2.3).

In case of the α3 subunit, the plasmid was digested using EcoRI restriction endonuclease

41


only (this is where the subunit cDNA was cloned originally into the multiple cloning site

of the vector). Next, the digested vector was ethanol precipitated (Section 2.1.6) and the

DNA pellet was resuspended in 8 µl 0.1 % (v/v) DEPC-treated water. Afterwards, vector

and subunit cDNA were religated using 1 µl T4 DNA ligase (1-3 U/µl; Promega) with 1 µl

10 × reaction buffer (300 mM Tris-HCl (pH 7.8 at 25°C), 100 mM MgCl2, 100 mM DTT

and 10 mM ATP) at 4°C overnight.

The pBS vector containing the α1 subunit was digested using HindIII and EcoRI restriction

endonucleases. The β2 and γ2L subunit cDNAs were digested using EcoRI and XbaI res-

triction endonucleases. 30 ng pcDNA3.1+ vector was digested for each subunit cDNA using

the respective restriction enzymes in a separate microfuge tube (Sections 2.1.4 and 2.2.3).

Digested plasmids containing the subunit cDNAs were run on a 2 % (w/v) agarose gel and

visualised by UV trans illumination (Section 2.1.4). Bands corresponding to α1, β2 and

γ2L subunit cDNAs were excised from the gels with a clean scalpel and purified using a

GFXTM PCR DNA and Gel Band Purification Kit (GE Healthcare; Section 2.1.8.2). Purified

subunit cDNAs as well as digested pcDNA3.1+ vector DNA were ethanol precipitated and

subsequently ligated using T4 DNA ligase (see above).

Ligation products were transformed into competent E. coli XL1 blue cells, amplified and

then isolated from the bacterial cultures. Subsequently, purified plasmid DNA was digested

and visualised on a 2 % (w/v) agarose gel. Plasmid DNAs containing an insert of the correct

size were sent for sequencing to confirm their identity and correct orientation in the vectors

(Sections 2.1.2 to 2.1.7).

2.3 Electrophysiological recordings in Xenopus oocytes

2.3.1 Making cRNA from cDNA

cRNA was made from cDNA using a T7 mMessage mMachine® Kit (Ambion) accor-

ding to the manufacturer’s protocol. GABAA receptor subunit cDNAs were present in the

pcDNA3.1+ (α3, α1, β2, γ2L and εT) or the pcDNA3 vector (ε); these vectors contain a

42


promoter sequence for T7 RNA polymerase. At first, the plasmid DNA was linearised by

cutting it after the end of the GABAA receptor subunit sequence. For this, 5 µg plasmid

DNA, containing the different GABAA receptor subunit cDNAs, respectively, was digested

using XbaI restriction endonuclease (8-12 U/µl, Promega) and 10 × buffer D (10 mM Tris-

HCl (pH 7.4), 300 mM NaCl, 0.1 mM EDTA, 1 mM DTT, 0.5 mg/ml BSA, 50 % glycerol)

in a total volume of 40 µl (Section 2.1.4). 2 µl of each linearised product was visualised on

a 1.5 % (w/v) agarose gel to confirm the completion of the digest, which was then termina-

ted by adding 1/20 volume 0.5 M EDTA (pH 8), 1/10 volume 3 M NaAc (pH 5.2) and 2 ×

volume 100 % (v/v) ethanol and incubated at -20°C for 30 minutes. Next, plasmid DNAs

were centrifuged for 15 minutes at 13,000 rpm and 4°C. Supernatants were removed and

DNA pellets resuspended in 0.1 % (v/v) DEPC-treated water to yield a final concentration of

0.5 µg/µl. Capped cRNA was transcribed in vitro using the T7 mMessage mMachine® Kit.

The transcription reaction was set up at room temperature (18-25°C) in the following order

and with a final volume of 20 µl:

Nuclease-free water 4 µl

2 × NTP/CAP (15 mM of each ATP, CTP, UTP, 3 mM GTP, 12 mM cap analog) 10 µl

10 × reaction buffer 2 µl

linearised cDNA template (0.5 µg/µl) 2 µl

T7 enzyme mix 2 µl

The mixture was incubated at 37°C for 70 minutes. Subsequently, 1 µl TURBO DNase

(2 U/µl) was added to each sample, which was incubated at 37°C for a further 15 minutes.

The resulting cRNA was recovered by performing a lithium chloride precipitation. 30 µl

nuclease-free water and 30 µl lithium chloride precipitation solution (7.5 M lithium chloride,

50 mM EDTA) were added to the cRNA and precipitated at -20°C for 40 minutes. The mi-

crofuge tube was then centrifuged at 13,000 rpm, 4°C for 15 minutes and the supernatant was

carefully removed. The cRNA pellet was washed with 70 % ethanol and once more centrifu-

ged for 15 minutes before the ethanol was removed and the cRNA pellet was resuspended in

an appropriate amount of nuclease-free water. Concentrations of the cRNAs were measured

43


using a UV spectrophotometer (DU530 or DU800, Beckman) as described in Section 2.1.5.

2 µl aliquots of cRNA samples at concentrations of 1 µg/µl or 2.5 µg/µl were stored at -80°C

until needed. Remaining cRNA was stored in 2.5 × volume 100 % (v/v) ethanol at -80°C.

2.3.2 Selection of Xenopus laevis oocytes

All following experiments using Xenopus laevis oocytes were performed in the laboratory

of Dr Ian Mellor (University of Nottingham). Oocytes that had been enzymatically separa-

ted and defolliculated with 2 mg/ml collagenase type 1A for 40-60 minutes were provided.

Healthy, large, round stage V and VI oocytes (1 to 1.2 mm diameter and 1.2 to 1.3 mm

in diameter, respectively) featuring a distinctive two-toned colour scheme (dark brown ani-

mal hemisphere and yellow vegetal hemisphere) were selected for GABAA receptor subunit

cRNA expression (Figure 2.2).

Figure 2.2: (A) Healthy and (B) impaired Xenopus laevis oocytesThe selection of healthy oocytes with a distinctive two-toned colour scheme and spherical or roundshape is vital for an effective expression of foreign proteins (from http://www.rothamsted.ac.uk/ppi/staff/TonyMiller/plantele_xenopus.html)

44


2.3.3 Injection of cRNA into Xenopus laevis oocytes

Oocytes were injected on the same day or one day after removal from the animal. Injection

pipettes were pulled from glass capillaries (No. 4878, World Precision Instruments, Inc.) in

one stage using a vertical pipette puller (Model 700C, David Kopf Instruments). Pipette tips

were shortened by gently tapping them against the lid of a sterile 35 mm Petri dish to a dia-

meter of approximately 15 µm. Subsequently, the pipette was backfilled with sterile-filtered

paraffin oil, using a hypodermic needle (25G) and 2 ml syringe, to seal the pipette from air.

It was then mounted onto a 50 nl microinjector (A203XVB, World Precision Instruments,

Inc.). The pipette was gradually filled with 4 µl cRNA mixture, containing one or more

cRNAs and sterile, DEPC-treated water. Selected oocytes were lined up in a 35 mm Petri

dish containing filter-sterilised modified Barth’s solution (96 mM NaCl, 2 mM KCl, 1.8 mM

CaCl2, 1 mM MgCl2, 5 mM HEPES, 2.5 mM pyruvic acid, 0.5 mM theophylline, 0.05 mg/

ml gentamycin; pH 7.5) and several silicone strips to keep oocytes separated and in place.

Oocytes were impaled with the injection pipette and 50 nl cRNA mixture was injected into

each oocyte at the following ratios: 1:1 for α3β2 receptors, 1:1:1 for α1/α3β2ε/εT receptors

(12.5 ng of each subunit per oocyte) and 1:1:10 for α3β2γ2L receptors (6.25 ng of α3 and

β2 subunit cRNA and 62.5 ng of γ2L subunit cRNA per oocyte). Oocytes were incubated

in modified Barth’s medium at 18°C for a minimum of 36 hours to allow GABAA receptor

expression. Oocytes, when injected with cRNAs encoding for α1 or α3, β2 and ε subunits,

were incubated in the presence of 100 µM Picrotoxin to block current through spontaneously

open channels. Oocytes were checked every day and dead oocytes were removed from the

medium. Barth’s medium was changed on the second day and on a regular basis afterwards.

2.3.4 Two-electrode voltage-clamp

Microelectrodes were pulled from thin wall borosilicate glass tubes (1.5 /1.12 OD/ID (mm);

Model no. TW150F-4, World Precision Instruments, Inc.) using a programmable, horizontal

pipette puller (P-97, Sutter Instruments) in two steps. Next, microelectrodes were bent in the

middle to about 150 degree using a spirit burner and backfilled with 3 M KCl using a modi-

fied 1 ml syringe to about two thirds full. To remove air bubbles in the microelectrode tip, it

45


was gently flicked until bubbles disappeared. It was then inserted into a microelectrode hol-

der mounted on a headstage (H2-RA, Axon Instruments) controlled by a micromanipulator

(Leitz). Microelectrodes had a resistance of 0.5 to 3 MΩ when submerged in the frog Ringer

solution (95 mM NaCl, 2 mM KCl, 2 mM CaCl2, 5 mM HEPES; pH 7.5). An oocyte was pla-

ced into the oocyte bath chamber (RC-3Z, Warner Instruments) and constantly perfused with

a frog Ringer solution at a flow rate of approximately 6-8 ml/min. Two microelectrodes were

impaled into the oocyte; one electrode recorded the voltage, the other injected the current (Fi-

gure 1.8, Chapter 1). Oocytes were clamped at a chosen holding potential, normally -60 mV,

using a GeneClamp 500 (Axon Instruments) or Axoclamp 2-A (Axon Instruments) amplifier.

Data were recorded to the hard disk of a PC using Strathclyde Electrophysiology Software

WinEDR2.9.2. (Dr J. Dempster, Department of Physiology and Pharmacology, University

of Strathclyde, UK). All drugs were bath applied by switching from frog Ringer solution

to drug solution using a ValveLink 8 or ValveBank 8 perfusion system (Automate Scienti-

fic). Oocytes were washed for at least four minutes between each drug application to avoid

receptor desensitisation. GABAA receptor agonists were applied alone and antagonists and

allosteric modulators were administered simultaneously with GABA at a concentration that

produced 20 % of the maximum response (EC20). GABA alone at EC20 concentrations ser-

ved as control. GABA, muscimol and zinc sulphate were prepared as 10 mM stock solutions

and picrotoxin was dissolved as a 100 µM stock solution in frog Ringer and diluted before

use. Flunitrazepam and etomidate were dissolved in 100 % (v/v) DMSO before dilution into

frog Ringer solution. The final solution did not exceed a DMSO level of 0.1 % (v/v), which

had no effect either alone or on GABA-elicited currents (Figure B.1 in Appendix B). All

experiments were performed at room temperature (18-25°C). A minimum of two batches of

oocytes were used per experiment.

2.3.5 Data analysis

Two-electrode voltage-clamp data were analysed using Strathclyde Electrophysiology Soft-

ware WinEDR V2.9.2 (Dr J. Dempster, Department of Physiology and Pharmacology, Uni-

versity of Strathclyde, UK) by measuring the amplitudes of peak-currents of recorded traces.

46


Concentration-response curves were fitted using GraphPad Prism 5.0 software (San Diego,

USA). Data were normalised and shown as percentage of the maximal response or percen-

tage of the control response, respectively, and fitted using a non-linear regression fit with

variable slope to the following equation:

response =

100/(1 + 10((logEC50−X)×n)) for agonists and positive modulating drugs

100/(1 + 10((log IC50−X)×n)) for antagonists(2.2)

Where X is the log of drug concentration, EC50/IC50 is the drug concentration evoking a

half-maximal response and n is the Hill coefficient. Data are given as means ± standard

error of the mean (SEM). EC50/IC50 values from different receptor subtypes were analysed

statistically using unpaired Student’s t-test and considered significant when p < 0.05.

2.4 Patch clamping mammalian cells

Glass coverslips (VWR) were cut into ~5.5 mm × 22 mm pieces using a diamond cutter.

Coverslip pieces were then sterilised by baking them at 180°C for 3 hours. Transiently

transfected HEK293 cells were reseeded at a low density onto sterile pieces of glass coverslip

in 35 mm Petri dishes (Nunc) 24 hours after transfection with GABAA receptor cDNAs (see

Sections 2.5.4 and 2.6.3). Recordings were performed using the whole-cell patch-clamp

configuration 24 to 48 hours after plating (Figure 1.9, Chapter 1).

Whole-cell recordings were performed in the laboratory of Dr Ian Mellor (University of Not-

tingham). Glass borosilicate capillaries (Model no. 1B150-4, World Precision Instruments,

Inc.) were pulled in two steps using a programmable, horizontal pipette puller (P-97, Sutter

Instruments). Pipettes were filled approximately two thirds with pipette solution (140 mM

CsCl, 1 mM CaCl2, 1 mM MgCl2, 11 mM EGTA and 5 mM HEPES; pH 7.2 adjusted with

CsOH) and air bubbles in the tip were removed by gently tapping the pipette. It was then

inserted into a pipette holder mounted on a CV 201AU headstage (Axon Instruments), which

was controlled by a micromanipulator (PCS - 5000 Series, Burleigh). Pipettes had a resis-

tance of ~5 to ~8 MΩ when lowered into the bath. A RC-27 bath chamber (Warner Ins-

47


truments) was placed on the stage of a microscope (Olympus). Coverslips with cells were

transferred into the chamber and continuously perfused with a mammalian Ringer solution

(135 mM NaCl, 5.4 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 5 mM HEPES and 10 mM Glu-

cose; adjusted to pH 7.4 with NaOH). Membrane currents were recorded using an Axopatch

200A amplifier (Axon Instruments). Cells were clamped at a holding potential of -50 mV.

Drugs were bath applied using a QMM micromanifold (ALA Scientific Instruments, Inc.)

by switching from mammalian Ringer solution to drug solution. 2 second pulses of agonists

or antagonists at specific concentrations were applied using a DAD-12 superfusion system

(ALA Scientific Instruments, Inc.). Cells were washed for at least 30 seconds between each

drug application. The amplifier, perfusion system and data recording to a PC were control-

led and synchronized by Strathclyde Electrophysiology Software WinWCP V2.9.2 (Dr J.

Dempster, Department of Physiology and Pharmacology, University of Strathclyde, UK).

All experiments were carried out at room temperature (18-25°C).

2.5 Cell culture

2.5.1 Culturing HEK293 cells

HEK293 cells were kindly provided by Prof. Anne Stephenson (The School of Pharmacy,

University of London, UK). Cells were grown in sterile-filtered Dulbecco’s modified Eagle’s

medium (DMEM), containing 4.5 g/l glucose and 4 mM L-glutamine, supplemented with

10 % (v/v) fetal bovine serum (FBS) and intermittently with 10 ml/l Penicillin-Streptomycin

(10,000 U/ml penicillin and 10 mg/ml streptomycin). Cells were maintained at 37°C in

a humidified atmosphere with 95 % (v/v) air and 5 % (v/v) carbon dioxide. Cells were

routinely grown in 75 cm2 or 25 cm2 tissue culture flasks (Sarstedt/Iwaki) and sub-cultured

twice a week (when 80-90 % confluent). For this, the growth medium was pipetted out

and cells were washed once with Dulbecco’s phosphate buffered saline (DPBS; 130 mg/l

CaCl2·2 H2O, 200 mg/l KCl, 200 mg/l KH2PO4, 100 mg/l MgCl2·6 H2O, 8 g/l NaCl, 2.16

g/l Na2HPO4·7 H2O) to remove traces of serum. Next, cells were incubated with trypsin/

EDTA (900 mg/l NaCl, 500 mg/l Trypsin, 200 mg/l Versene (EDTA)·2 Na·2 H2O) at 37°C

48


until they detached from the plastic flask and fresh growth media was added to quench the

action of the trypsin/EDTA. Cells were centrifuged at 300 rcf, 22°C for 5 minutes (Models

5415R, Eppendorf or Universal 320R, Hettich); the supernatant was discarded and the cell

pellet resuspended in 1 ml fresh medium by pipetting the cells gently up and down several

times. Afterwards, a suitable quantity of cells was transferred into a tissue culture flask

containing fresh growth medium and incubated as stated above. Cells were cultured for

about 20 passages before a new stock was regenerated. All reagents were purchased from

Lonza.

2.5.2 Long term storage of HEK293 cells in liquid nitrogen

For cryopreservation, HEK293 cells of a low passage number were grown and harvested as

described above (Section 2.5.1). Cell pellets were carefully resuspended with a pipette until

cells were free of clusters. Next, cells were counted using a hemocytometer with Neubauer

improved grid pattern (C-Chip (DHC-N01), Incyto) by transferring 20 µl resuspended cells

to a microfuge tube containing an appropriate volume of growth medium, depending on the

size of the tissue culture vessel where cells grew in. Cells were further diluted 1:1 (final

concentration of 4 to 6 × 106 cells per ml) using a Trypan blue solution (0.4 % in 0.81 %

NaCl and 0.06 % K2HPO4, Sigma), which stained non-viable cells blue. 10 µl of the mixture

were loaded into the counting chamber and viable cells were counted in 5 large squares. The

number of viable cells per ml was calculated as follows:

no. of cells counted5

× dilution factor× 104 = cells/ml (2.3)

Afterwards, cells were pelleted again and resuspended in sufficient freezing medium (DMEM

supplemented with 10 % (v/v) FBS and 5 % (v/v) DMSO) to get a final concentration of

2 × 106 cells per ml. 1 ml of cell suspension was pipetted into each cryovial and frozen at

-80°C for 24 hours before vials were transferred into liquid nitrogen for long term storage.

49


2.5.3 Thawing cells stored in liquid nitrogen

Cryovials of HEK293 cells from frozen stock were thawed quickly by incubating the vials

at 37°C and transferring the cells to 20 ml prewarmed growth medium. Next, the cell sus-

pension was centrifuged at 300 rcf, 22°C for 5 minutes and cell pellets were resuspended in

1 ml fresh growth medium before transferring them to a tissue culture flask containing fresh

growth medium and incubating them at 37°C, 95 % (v/v) air and 5 % (v/v) carbon dioxide

until 80 to 90 % confluent (Section 2.5.1).

2.5.4 Transient transfection of HEK293 cells

Cells were seeded into different tissue culture vessels (depending on the number of cells

needed) in DMEM containing 10 % FBS, but no antibiotics, and grown for 24 hours as

described in Section 2.5.1 and Table 2.3.

Table 2.3: Transient transfection using TransPassTM D2 in different culture vessels

Culture vessel No. of cellsseeded

Volume ofmedium

total amount plas-mid DNA in DMEM

TransPassTM

D224-well plate 1.5 × 105 250 µl 0.7 µg in 50 µl 1.75 µl6-well plate 7.2 × 105 1 ml 3 µg in 250 µl 8.75 µlT25 flask 1.85 × 106 3.2 ml 9 µg in 600 µl 22 µlT75 flask 5.55 × 106 9.6 ml 27 µg in 1800 µl 66 µl

After 24 hours, cells were 70-80 % confluent and transfected using TransPassTM D2 trans-

fection reagent (New England Biolabs) according to the manufacturer’s protocol. Briefly,

plasmid DNA was pipetted into DMEM as in Table 2.3, when transfecting several different

DNAs the quantity was divided equally or at certain ratios. Next, 22 µl TransPassTM D2 was

added, mixed carefully and incubated for 20 to 30 minutes at room temperature (18-25°C)

to facilitate formation of tranfection complexes. Subsequently, the mixture was added to the

cells and incubated for 4 to 6 hours as described before (Section 2.5.1). Medium containing

excess transfection complexes was then replaced by fresh growth medium in order to prevent

toxic effects to the cells. Transient transfection was optimised using pRK5-Clomeleon as a

reporter protein (Section 2.5.4.1).

50


2.5.4.1 Evaluation of transient transfection efficiency

Transient transfection efficiency was optimised by transfecting HEK293 cells with pRK5-

Clomeleon, which served as fluorescence marker of transfected cells. Cells were seeded into

and transfected in 24-well plates as described in Section 2.5.4, but varying the number of

cells as well as the plasmid DNA to TransPassTM D2 transfection reagent ratio. Approxi-

mately 24 hours after the transfection, cells were removed from the wells by using trypsin/

EDTA (as described in 2.5.1). Transfection efficiency was examined by placing cells in solu-

tion onto a glass microscope slide and visualising them using an Olympus BX51 fluorescence

microscope. A minimum of 1000 cells were counted per experiment using CellF software

(Olympus) in order to determine the ratio of transfected (fluorescent) cells versus untrans-

fected cells. The conditions resulting in highest transfection efficiency after performing each

experiment in duplicates were used for all further transient transfection experiments (Section

2.5.4, Table 2.3).

2.5.5 Reseeding cells into appropriate tissue culture plasticware

Typically, cells had to be reseeded into cell culture plastic ware suited for respective expe-

riments 24 hours after transient transfection. For this, cells were washed once using DPBS

and subsequently detached from the surface by adding trypsin/EDTA and incubating them at

37°C until dissociated from the plastic. Sufficient growth medium was added to the cells to

quench the trypsin/EDTA. Cells were then centrifuged at 300 rcf for 5 minutes at room tem-

perature (18-25°C). Cell pellets were resuspended and counted (Section 2.5.2); the required

number of cells was plated into microtiter plates or coverslips placed in 35 mm Petri dishes

and incubated further until usage in respective experiments (see below).

51


2.6 Assay development to screen compound libraries

2.6.1 A cell-based assay using the chloride-ion indicator Clomeleon

2.6.1.1 Excitation and emission spectra of Clomeleon

Sequencing the pRK5-Clomeleon plasmid revealed a discrepancy of one amino acid with the

sequence data provided by Dr Thomas Kuner. Therefore, the excitation and emission spectra

of Clomeleon were investigated using a spectrofluorometer (QuantaMaster, Photon Techno-

logy International,Inc.). HEK293 cells were seeded into 24-well plates and transfected 24

hours later with GABAA receptor subunit cDNA (α3 and β2 and γ2L) and pRK5-Clomeleon

at a ratio of 1:1:1:1 (Section 2.5.4). On the next day, cells were reseeded onto sterile covers-

lips (10 × 24 mm, Chance Propper Ltd.) placed in 35 mm dishes (Nunc) at a cell density of

7 to 10 × 105 in 500 µl medium. Cells were incubated as described above (Scetion 2.5.1)

until the following day.

Excitation and emission screens were performed using a spectrofluorometer (QuantaMaster,

Photon Technology International,Inc.). Cells on a coverslip were immersed into a fluorime-

ter cuvette (Sigma) containing mammalian Ringer solution (Section 2.4), which was subse-

quently fitted into the spectrofluorometer. The emission scan was performed at wavelengths

of 450 nm to 650 nm, with an excitation of 440 nm; the excitation scan was carried out from

350 nm to 485 nm with the emission being set to 527 nm. Both were performed with cells

transfected with Clomeleon alone or Clomeleon and α3β2γ2L GABAA receptors. Additio-

nally, untransfected HEK293 cells, a cuvette with mammalian Ringer solution and a cuvette

alone were scanned as controls. Experiments were performed at 37°C.

2.6.1.2 Evaluating a screening assay using Clomeleon

HEK293 cells were seeded into 25 cm2 tissue culture flasks and 24 hours later transfected

with GABAA receptor subunit cDNAs (α3 and β2 or α3, β2 and γ2L) and pRK5-Clomeleon

at ratios of 1:1:1 or 1:1:1:1, respectively (Section 2.5.4). On the following day, cells were pla-

ted into Poly-D-Lysine coated, black 96-well plates with clear, flat bottom (BD Biosciences)

52


at cell densities of 1.5 × 105 and 2 × 105 cells per well in 200 µl medium. Poly-D-Lysine

coating enhanced cell growth and attachment to the wells and black polystyrene plates were

used to lower background noise and well-to-well crosstalk. Cells were incubated for another

day before performing the assay using a FLUOstar Optima microplate reader (BMG Lab-

tech) fitted with one reagent injector, a 430 ± 10 nm excitation and 480 ± 10 nm and 520-P

emission filters.

On the day of the assay, the growth medium was removed from each well and replaced with

90 µl prewarmed mammalian Ringer solution (Section 2.4) per well. Cells were then incuba-

ted for a further one hour (as described in 2.5.1) before placing the plate into the FLUOstar

Optima plate reader, which was set to excite and read Clomeleon emission from the bottom

of the plate. Baseline fluorescence was measured for 20 seconds (0.2 second intervals) for

both Clomeleon emission wavelengths (480 nm and 520 nm), together as well as separa-

tely. Next, 10 µl GABA (1 mM) in mammalian Ringer solution was injected into each well

(100 µM final GABA concentration). To intermix the added solution, the plate was shaken

for 1 second before the fluorescence signal was continued to be read for a further 30 seconds

at 0.2 second intervals. Experiments were performed at 37°C and repeated twice. Data were

analysed by normalising averaged baseline fluorescence signal to 100 % and fluorescence

signals recorded after GABA solution injection were shown as percentage of baseline fluo-

rescence. Data are given as means ± SEM.

2.6.2 Setting up an iodide-flux assay

The assay was set up as described in Tang & Wildey (2004) and Tang & Wildey (2006).

HEK293 cells were seeded into 25 cm2 tissue culture flasks and transfected 24 hours later

with GABAA receptor subunit cDNAs (α3, β2 and γ2L, ratio 1:1:1; Section 2.5.4). On

the following day, cells were plated into Poly-D-Lysine coated 96-well microplates with

flat bottom (BD Biosciences) at cell densities of 25 × 103 cells per well in 100 µl growth

medium. At a later date, cells were tested at lower densities of 12.5 × 103 and 17.5 × 103

cells per well. Cells were incubated for another day before proceeding with the iodide-flux

assay. Untransfected HEK293 cells served as negative controls.

53


DMEM was removed from the wells and cells were loaded with 100 or 200 µl filter-sterilised,

prewarmed iodine-loading buffer (150 mM NaI, 2 mM CaCl2, 0.8 mM NaH2PO4, 1 mM

MgCl2, 5 mM KI, 2 % FBS; adjusted to pH 7.4 with NaOH). Subsequently, cells were

incubated for a further 4 hours at 37°C, 95 % (v/v) air, 5 % (v/v) carbon dioxide to allow cell

loading. Cells were then washed 3 to 4 times with 200 µl DPBS to remove excess iodide,

before adding 100 µl DPBS containing increasing concentrations of GABA and incubated

for a further 5 minutes at room temperature (18-25°C). Next, the DPBS containing GABA

and the ’effluxed’ iodide was transferred to a fresh plate. Cells on the first plate were lysed by

adding 100 µl lysis buffer (1 % (v/v) triton X-100). Colorimetric detection was accomplished

on both plates (lysed cells as well as DPBS) by adding to each 100 µl detection buffer II

(2 mM Ce(NH4)4(SO4)4·2H2O, 5.2 % (v/v) fuming H2SO4 in distilled H2O) followed by

100 µl detection buffer I (1.25 % (v/v) NH4OH solution, 100 mM As2O3, 3.2 % (v/v) fuming

H2SO4, 467.38 mM NH4Cl). Subsequently, the plate was incubated at room temperature for

15 minutes. Next, it was inserted into a microplate reader (ASYS, Expert 96), shaken for 2

seconds to evenly disperse the liquid and the OD405 was read.

The results obtained were rather random, therefore, several assay modifications and optimi-

sations were tested. Since the OD405 readings were relatively low, not only less cells were

seeded, but also it was examined whether the results could be improved by diluting the lysed

cells and DPBS (1:2 and 1:5) before performing the colorimetric assay. In order to test func-

tionality of GABAA receptors, cells were incubated in DPBS with 50 µM bicuculline and 1

or 100 µM GABA or in DPBS containing 50 µM picrotoxin and 1 or 100 µM GABA as well

as 1 or 100 µM GABA and DPBS alone. Untransfected cells served as controls. Different

agonist incubation times of 10 seconds up to 5 minutes (DPBS containing 10 µM GABA)

were also investigated to optimise efflux. In every assay for every test variable 8 replicates

were prepared per plate. Different setups were repeated at least 2 times.

A standard curve was generated by adding 100 µl per well in a 96-well plate of ascending

concentrations of NaI (3 nM to 300 µM) as well as distilled water as negative control. Af-

terwards, 100 µl detection buffer II and 100 µl detection buffer I were added into each well,

contents were mixed and incubated at room temperature for 10 to 70 minutes before reading

54


the OD405 on the microplate reader. 8 standard replicates per NaI concentration were pre-

pared and the experiment was repeated 3 times on different days. Data points were averaged.

The assay detection limit was defined as minimal NaI concentration that fulfils the following

argument (Tang & Wildey, 2004):

OD405sample −OD405control > 3SDcontrol (2.4)

Where OD405sample is the OD405 of the sample containing NaI, OD405control is the OD405

of the control sample that contains no NaI and SDcontrol is the standard deviation (SD) of the

control.

OD405 readings of the iodide-flux assay were averaged and the SD was calculated. Values

were analysed statistically using unpaired Student’s t-test and considered significant when

p < 0.05. In some cases, channel conductivity (% activity) was calculated as follows (Tang

& Wildey, 2004):

(OD405sample −OD405low)/(OD405high −OD405low)× 100 (2.5)

Where OD405sample was the reading of the wells incubated with different GABA concentra-

tions, OD405high was the reading of the cells treated with the highest concentration of GABA

and OD405low was measured from cells incubated with distilled water instead.

2.6.3 Setup and optimisation of a mutant YFP-based assay

HEK293 cells were seeded into 25 cm2 or 75 cm2 tissue culture flasks. On the next day,

when 70 - 80 % confluent, cells were transfected with GABAA receptor subunit cDNAs and

the pEYFP H148Q/I152L vector (as described in 2.5.4). GABAA receptor subunit cDNAs to

pEYFP H148Q/I152L vector ratios were as listed below:

55


α1 : β2 : pEYFP H148Q/I152L → 1:1:2

α3 : β2 : pEYFP H148Q/I152L → 1:1:2

α1 : β2 : ε : pEYFP H148Q/I152L → 1:1:1:2

α3 : β2 : ε : pEYFP H148Q/I152L → 1:1:1:2

α3 : β2 : γ2L : pEYFP H148Q/I152L → 1:1:1:2

Untransfected cells served as control cells. On the following day, cells were plated into Poly-

D-Lysine coated, black-sided 96-well plates with clear, flat bottom (BD Biosciences) and

incubated for a minimum of a further 24 hours. For assay optimisation, different numbers

of cells were seeded per well (in 200 µl growth medium). Approximately 1.5 × 105 cells

per well (in 200 µl medium) were decided to give the best results and used in all further

experiments.

On the day of the assay, one hour prior to the start of an experiment, growth medium was

removed from the wells and replaced by 50 µl prewarmed NaCl bathing solution per well

(140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, 10 mM Glucose;

pH 7.3 using NaOH). Following incubation, the plate was placed into a FLUOstar Optima

microplate reader (BMG Labtech) fitted with two reagent injectors, a 500 ± 10 nm excita-

tion and a 520-P emission filter. Assays were performed at 37°C. The mutant YFP-H148Q/

I152L was excited and its emission was measured from the bottom of the plate; each well

was scanned orbitally at a radius of 3 mm. Baseline fluorescence was recorded for 5 seconds

(one second intervals) before injection of 150 µl NaI test solution only (140 mM NaI, 5 mM

KCl, 2 mM CaCl2, 1 mM MgCl2, 10 mM HEPES, 10 mM Glucose; pH 7.3 using NaOH)

or NaI test solution supplemented with GABA or picrotoxin at specific concentrations (in-

jection speed 100 µl/sec). After solution was injected, the fluorescence signal was recorded

at 0.5 second intervals for a further 30 seconds. Experiments were repeated 3 times, each

with at least 3 wells on the same plate and at least one of the replicates was performed on a

separate plate.

As it became clear that the assay was functional, GABA concentration-response relation

curves were determined for α1β2, α1β2ε and α3β2γ2 receptors by injecting ascending

56


GABA concentrations (in NaI test solution) ranging from 30 nM to 3 mM. By adding NaI test

solution containing flunitrazepam (1 µM final concentration) and an EC30 concentration of

GABA to cells expressing α3β2γ2 receptors it was demonstrated that the assay was suitable

to detect positive modulation of the receptors. Flunitrazepam stock solution was prepared

by dissolving it in 100 % (v/v) DMSO. Before the assay, flunitrazepam was diluted into NaI

test solution. The final solution did not exceed a DMSO level of 0.5 % (v/v), which had no

effect on GABA-elicited currents (Figure B.2 in Appendix B).

2.6.3.1 Data analysis

Data were analysed by subtracting the averaged fluorescence measured from non-transfected

control cells (baseline fluorescence and after addition of NaI test solution) resulting in the net

fluorescence signal. This was then normalised by setting the averaged baseline fluorescence

to 100 % and fluorescence changes (after adding solutions) were calculated as percentage of

the baseline fluorescence. The percentage of fluorescence quench was calculated as follows

(Gilbert et al., 2009a):

% fluorescence quench = (1− (F final/F BF))× 100 (2.6)

FBF is the averaged baseline fluorescence signal and Ffinal is the averaged final fluorescence

signal. Values were presented as means ± SEM. To assess the assays quality without test

compounds the Z’factor was calculated as described below (Zhang et al., 1999):

Z ′ = 1− (3σc+ + 3σc-)

|µc+ − µc-|(2.7)

σc+ or σc- represent the SD of positive and negative controls, µc+ or µc- states the means

of positive and negative controls. Positive controls were readings from cells treated with a

high GABA concentration (100 µM), for negative controls NaI test solution only was applied

to the cells. Concentration-response curves were fitted using GraphPad Prism 5.0 software

(San Diego, USA). Data were shown as percentage of baseline fluorescence and fitted using

57


a non-linear regression fit with variable slope according to the following equation:

response =

bottom + (top + bottom)/(1 + 10((logEC50−X)×n)) (2.8)

Where X is the log of drug concentration, EC50 is the drug concentration evoking a half-

maximal response, n is the Hill coefficient and the response starts at bottom and goes to top.

Data are given as means ± SEM.

2.7 Compound screening using mutant YFP-H148Q/I152L

2.7.1 Chemical compound library

800 compounds covering a broad range of structural motifs and biological space were acqui-

red from the Maybridge HitFinderTM collection of screening compounds (Maybridge Ltd.).

Compound quality was guaranteed to have a minimum of 90 % purity. Compounds were

arrayed individually in ten separate 96-well plates (80 compounds per plate) as dry films of

1 µmol per compound and stored at -80°C until usage.

Compounds were resuspended in 100 µl DMSO per well. Plates were placed on an orbi-

tal shaker (Stuart) and left until compounds were dissolved. Subsequently, 20 µl aliquot

plates were prepared. All plates were sealed using parafilm and stored at -20°C until needed.

Concentrations of these compound stock solutions ranged from ~10.1 to ~11.9 mM.

2.7.2 Screening assay

HEK293 cells were seeded in 75 cm2 flasks, transfected with α1, β2, ε or α3, β2, γ2L

GABAA receptor subunit cDNAs and YFP-H148Q/I152L cDNA and reseeded into Poly-D-

Lysine coated 96-well plates (Sections 2.5.4, 2.6.3). On the next day, the assay was perfor-

med as described in Section 2.6.3. Application of 100 µM GABA, 1 mM picrotoxin with an

EC30 GABA concentration as well as EC30 GABA with or without 0.5 % DMSO (all were

applied in NaI test solution) served as controls.

58


1 µl of compound stock solution was transferred manually, using a multichannel pipette, to

each well of the microtiter plate containing 50 µl NaCl bath solution. Next, the plate was

returned to the plate reader and shaken for 10 seconds to intermix the added compound.

150 µl of low concentration GABA (EC30) in NaI test solution was injected into the first well

and the fluorescence signal was read for 30 seconds (0.5 second intervals), before GABA in

NaI test solution was added to the next well to measure the fluorescence signal for 30 seconds

until all compounds were screened for their possible modulating or antagonist activity. Time

elapsed between the addition of a chemical compound to the wells until the injection of NaI

test solution and subsequent fluorescence measurement was between approximately 5 and

60 minutes. In later assays this time was reduced to shorter periods.

The initial screening was performed with relatively high compound concentrations (final

concentrations were between 50.3 and 59.29 µM). Due to the high number of ’hit’ com-

pounds, which enhanced or inhibited submaximal GABA-evoked receptor activity, the scree-

nings were repeated several times using gradually decreased compound concentrations (com-

pound solutions were further diluted using DMSO). During these preliminary screens, each

compound was tested in singlets until the number of ’hit’ compounds was whittled down suf-

ficiently. The remaining ’hit’ compounds were then tested in at least three wells on the same

plate, repeated at least three times. Compounds applied with NaI test solution only were

examined for possible agonist activity of ’hit’ compounds and cells transfected with YFP-

H148Q/I152L cDNA only served as control cells to rule out compound activity via receptors

expressed endogenously in the cells. Data were analysed as in Section 2.6.3 described.

2.7.2.1 Data analysis

The effect of test compounds was compared to the EC30 GABA-control response and analy-

sed statistically using Student’s t-test. Results were considered significant when p < 0.05.

2.7.3 Further testing of ’hit’ compounds

A selection of ’hit’ compounds which exhibited receptor modulation or inhibition during the

preliminary screening assay was further tested in Xenopus laevis oocytes expressing α1β2ε

59


or α3β2γ2 receptors using the two-electrode voltage-clamp technique (Section 2.3). Test

compounds were dissolved in DMSO as 100 mM stock solutions before dilution into frog

Ringer solution. The final solution did not exceed a DMSO level of 0.1 % (v/v), which had

no effect either alone or on GABA-elicited currents (Figure B.1, Appendix B).

60

Chapter 3

Pharmacological characterisation of the

α3β2ε GABAA receptor subtype

3.1 Introduction

3.1.1 The GABAA receptor ε subunit

The GABAA receptor ε subunit belongs to a minor receptor population with unknown func-

tion (Olsen & Sieghart, 2009). Its expression in the brain is restricted to specific areas, mainly

areas in the forebrain, such as the hypothalamus, as well as nuclei containing major modula-

tory systems of the brain, including cholinergic cell groups (Moragues et al., 2000; Sieghart

& Sperk, 2002). The latter also express the α3 subunit. The ε subunit is clustered together

with the α3 and θ subunits on chromosome Xq28 of the human genome (Wilke et al., 1997)

and it has been shown that these subunits are co-expressed in several areas of the brain (Mo-

ragues et al., 2000; Sieghart & Sperk, 2002). It is therefore likely that they form functional

receptor channels in some brain areas. Heterologous expression studies have shown that the

ε subunit can replace the γ subunit to form functional GABAA receptors, when co-expressed

with α and β subunits (Davies et al., 1997; Whiting et al., 1997). These receptors respond to

a wide range of allosteric modulators, but exhibit unusual characteristics compared to other

known native GABAA receptors, such as increased desensitisation rate, high agonist sensiti-

vity and spontaneous channel opening (see Section 1.2.5.1, Chapter 1; Whiting et al., 1997;

61

Chapter 3. Characterisation of the α3β2ε GABAA receptor

Neelands et al., 1999).

The subunit compositions of ε-containing receptors in vivo and their precise functions are

largely unknown. Previously, the pharmacological profile of ε-subunit-containing receptors

has been investigated in heterologous expression systems in several different subunit combi-

nations. Nonetheless, despite co-expression of α3 and ε subunits in a number of brain areas,

only one study examined pharmacological properties of recombinant receptors containing

these subunits in conjunction with the β1 subunit (Ranna et al., 2006). However, the β2 su-

bunit is more abundant and the most widely expressed β subunit in the brain. In several areas

of the brain, α3 and ε subunits are co-expressed with β2 subunits, for example in hypotha-

lamus and amygdala (Zhang et al., 1991; Moragues et al., 2000; Pirker et al., 2000; Sieghart

& Sperk, 2002) and it is likely that these subunits may form functional receptors in some of

those areas. In order to learn more about the role of the ε subunit it is, therefore, important

to know the pharmacological profile of α3β2ε receptors.

In this study, pharmacological properties of the α3β2ε receptor subtype were investigated

and compared with α3β2 and α3β2γ2 receptors. The initial plan was to express cDNA

encoding the different subunits in HEK293 cells, which were also used later in the cell-based

screening assay (see Chapters 4 and 5). In this way, not only the pharmacological profile

of this novel subunit combination could be examined, but the correct subunit assembly in

HEK293 cells could be investigated too (which would be advantageous for setting up the

screening assay). However, problems with the electrophysiological rig at Nottingham Trent

University caused a slight change in the project and electrophysiological recordings were

instead performed in Xenopus laevis oocytes microinjected with cRNA in the laboratory of

Dr Ian Mellor at the University of Nottingham.

3.1.2 ε subunit splice variants

As mentioned in Chapter 1 (Section 1.2.2), subunit splicing adds even more complexity and

diversity to the GABAA receptor family. It has been shown, that several of the GABAA

receptor subunits exist in different splice variants, including α2, α4, α5, α6, β2, β3, γ2, γ3

62


and ε subunits (Simon et al., 2004). Often, there are two different splice forms, a long and a

short version of the protein, for example the β2 and γ2 subunits exist in both forms. In the

case of the γ2 subunit, the long form exhibits an eighth amino acid insertion into the large

intracellular loop between TM3 and TM4, introducing a site for phosphorylation by protein

kinase C (PKC; Whiting et al., 1990).

The ε subunit gene consists of nine exons (Sinkkonen et al., 2000; Simon et al., 2004) and

several mRNA splice variants have been demonstrated in brain as well as in several other

peripheral tissues. Four splice forms were described by Wilke et al. (1997): (1) the full

length, functional transcript, which is expressed in brain, but also found in heart and spinal

cord, (2) a version where the first three exons are spliced out and (3) a variant like 2, but with

an additional deletion in the centre of exon four; both variants are expressed in peripheral

tissues, but not found in the brain, and (4) a deletion of exon one only. Moreover, Sinkkonen

et al. (2000) described a 483 amino acid long insertion encoding a Pro/Glx motif, which

was identified exclusively in rodents. However, Moragues et al. (2000) identified a rodent

ε version lacking this long insertion, and, therefore, resembling the human sequence to a

greater extent. It was demonstrated that both versions are expressed in the brain as well as

mRNA of a further, truncated version (described below, Figure 3.2; Kasparov et al., 2001).

Davies et al. (2002) investigated the functionality of the long rodent splice variant by expres-

sing it in HEK293 cells. Functional properties as well as the transcripts expression pattern

were examined, but failed to show its incorporation into functional channels. However, since

there was no premature stop codon and the sequence is in frame, it was concluded that there

may be a different function of this long transcript. The splice properties of the ε subunit

are relatively unusual compared to other subunits, since splicing can be introduced at several

different positions. Many of the truncated subunit transcripts are found in several different

tissues, where they may exhibit unknown functions, and, also, several splice variants exist in

the brain.

In this project, the truncated ε subunit version (εT) found in brain (Kasparov et al., 2001)

was generated and functionally examined. This variant possesses a 26 amino acid deletion,

starting in the centre of TM2 until the end of the extracellular loop between TM2 and TM3

63


(Figures 3.1 and 3.2). In order to find further clues about the role of this unusual subunit,

it was investigated whether the truncated version forms functional channels in vitro when

co-expressed with the α3 and β2 subunits.

TM1 TM2

Rat ε IVFQNYIPSSVTTMLSWVSFWIKIEAAAARASVGVSSVLTMATLGRat εT IVFQNYIPSSVTTMLSWVSFWIKIEAAAARASV............

TM2 TM3

Rat ε TFSRKNFPRVSYLTALDFYIAICFVLCFCTLLEFTVRat εT ..............ALDFYIAICFVLCFCTLLEFTV

Figure 3.1: Alignment of the partial rat GABAA receptor ε subunit amino acid sequence(single-letter code) in the region of TM1 to TM3.The truncated version (εT; accession number AF255385) shows a 26 amino acid deletion fromthe middle of TM2 to the start of TM3, compared to the full length version (accession numberAF255612).

64


extra

intra

–NH2•M•L•P•K•V•L•L•M•L•L•N•M•F•L•A•L•Q•W•R•V•G•P•H•I•K•L•E•N•K•P•P•A•Q•D•K•V•V•F•G•P•Q•P•S•G•K•K•L•P•A•R•E•T

•E•L•T•A•D•H•T•T•E•R•P•R•G•K•L•T•R•A•S•Q•I•L•N•T•I•L•S•N•Y•D•H•K•L•R•P•S•I•G•E•K•P•T•V•V

•T•V•K•V•F•V•N•S•L•G•P•I•S•I•L•D•M•E•Y•S•I•D•I•I•F•Y•Q•T•W•Y•D•E•R•L•R•Y•N•D•T•F•E•T•L•I

•L•H•G•N•V•V•S•Q•L•W•I•P•D•T•F•F•R•N•S•K•R•T•Q•E•Y•D•I•T•I•P•N•Q•M•A•L•I•H•K•D•G•K•V•L•Y

•T•V•R•M•T•I•D•A•R•C•S•L•H•M•L•N•F•P•M•D•S•H•S•C•P•L•S•F•S•S•F•S•Y•D•E•H•E•M•I•Y•K•W•E•N

•F•K•L•K•I•D•A•K•N•T•W•K•L•L•E•F•D•F•T•G•V•N•N•K•T•E•I•I•S•T•P•V•G•D•F•M•V•M•T•F•F•F•N•V•S•R•R•F•G•F•I•V•F•Q •N•Y•I •P•S•S •V•T•T •M•L•S •W•V•S•F•W•I•K•I•E•A•A•A•A•R•A•S•V•V•

S•S•V•L•T•M•A•T•L•G•F•S•R•K•N•F•P•

R•V•S•Y•L•A•L•D•F •Y•I•A •I•C•F•V •L•C•F•C •T•L•L •E•F•T•V•L•N•F•L•T•Y•N•N•I•E•R•Q•A•S•P•K•F•Y•Q•F•P•T•N•S•R•A•N•A•R•T•R•A•R•A•R•T•R•A•R•A•R•A•R•A•R•Q•Q•Q•E•V•F•V•C•E•I•V•T•Y•E•E•N•A•E•E•G•Y•Q•W•S•P•R•S•R•R•P•Q•C•P•W•R•R•C•G•R•S•Y•V•C•F•R•V•L•R•K•Y•F•C•M•V•P•G•C•E•G•N•S•W•Q•R•G•R•I•C•I•H•V•Y•R•L•D•N•Y•S•R•V•L•F•P•I•T•F•F•

F•F•N•V•V•Y•W•V•I•C•L•N•L•X –COOH

312•G•T

337•TI II III IV

Figure 3.2: Structure of the ε subunit; amino acid sequence is displayed in single-letter code.Orange circles represent the 26 amino acid deletion in the truncated subunit version.

65


3.2 Results

3.2.1 Agonist sensitivity

α3β2, α3β2ε and α3β2γ2 GABAA receptor subtypes were expressed in Xenopus laevis

oocytes and examined using the two-electrode voltage-clamp technique. To characterise the

pharmacological and biophysical properties of the receptors, they were first examined for

their sensitivity to the agonists GABA and muscimol. In all cases both agonists elicited

concentration-dependent inward currents (see Figure 3.3 for GABA-evoked traces), whereas

no currents were detected upon the application of 1 mM GABA in oocytes expressing α3,

β2, γ2 and ε subunits alone (n = 10 for α3, β2 and ε subunits and n = 16 for γ2 subunits).

It is noteworthy that maximal currents elicited by GABA were relatively small in α3β2ε

receptors compared to α3β2 and α3β2γ2 receptors (Figure 3.3).

GABA GABA GABA

GABA GABA

500 nA

50 sec

GABA

GABA GABA GABA GABA

50 sec

250 nA

GABA

50 sec500 nA

GABAA

B

C

Figure 3.3: Sample current traces evoked by ascending GABA concentrations in oocytesexpressing α3β2 (A), α3β2ε (B) and α3β2γ2 subunit receptors (C).Sample currents were elicited by GABA concentrations of 1 µM, 10 µM, 100 µM and 1 mM. Theholding potential was -60 mV.

Concentration-response curves of GABA as well as muscimol for the different receptor sub-

types are depicted in Figures 3.4 and 3.5 and EC50 and Hill slope data obtained are summa-

66


rised in Table 3.1. EC50 values differed significantly between the receptor subtypes. α3β2ε

receptors were the most sensitive receptors to both agonists, whereas the α3β2γ2 receptor

subtype was the least sensitive. In addition to the high agonist sensitivity, the α3β2ε receptor

subtype also exhibited a moderate Hill slope for both agonists, whereas Hill slopes for the

other two receptor types were more steep (Table 3.1).

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1

0

25

50

75

100

α3β2α3β2εα3β2γ2

log [GABA, M]

% m

axim

um r

espo

nse

Figure 3.4: GABA concentration-responsecurves for GABAA receptors expressed inXenopus laevis oocytes.GABA concentration-response curves of α3β2(n = 11), α3β2ε (n = 12) and α3β2γ2 recep-tor subtypes (n = 10). The data correspond tomeans ± SEM; n states the number of oocytesexamined. EC50 values and Hill coefficients arepresented in Table 3.1. Curves were fitted usinga non-linear regression fit with variable slopesas described in Section 2.3.5.

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1

0

25

50

75

100


log [Muscimol, M]%

max

imum

res

pons

e

Figure 3.5: Muscimol activation ofGABAA receptors expressed in Xenopuslaevis oocytes.Muscimol concentration-response curves ofα3β2 (n = 6), α3β2ε (n = 6) and α3β2γ2receptor subtypes (n = 7). Data correspond tomeans ± SEM; n represents the number of oo-cytes tested. EC50 values and Hill coefficientsare presented in Table 3.1. Curves were fittedusing a non-linear regression fit with variableslopes as described in Section 2.3.5.

Table 3.1: Effects of GABA and muscimol on different GABAA receptor subtypes

GABA MuscimolEC50 (µM) Hill slope EC50 (µM) Hill slope

α3β2 3.30 ± 1.17 (n = 11) 0.82 ± 0.18 0.53 ± 1.65 (n = 6) 1.24 ± 0.14α3β2ε 0.23 ± 0.89 (n = 12) 0.44 ± 0.13 0.06 ± 1.02 (n = 6) 0.54 ± 0.06α3β2γ2 39.57 ± 1.35 (n = 10) 0.77 ± 0.10 2.77 ± 1.37 (n = 7) 1.08 ± 0.18

Data represent mean values (± SEM). EC50 and Hill slope values were obtained by non-linear regres-sion fit with variable slopes as described in Section 2.3.5. n represents the number of oocytes used perexperiment. GABA EC50 values were significantly different in oocytes expressing α3β2ε (p < 0.001)and α3β2γ2 receptors (p < 0.0001), when compared with values obtained for α3β2 receptors (un-paired t-tests). Likewise, muscimol EC50 values differed significantly in oocytes expressing α3β2ε(p < 0.0001) and α3β2γ2 receptors (p < 0.01; compared to α3β2 receptors).

67


3.2.2 Current-voltage relationships

Current-voltage relations for the α3β2, α3β2ε and α3β2γ2 GABAA receptor subtypes were

determined by administrating submaximal GABA concentrations (EC20) at different holding

potentials (ranging from -90 to -10 mV; Figure 3.6). The reversal potential for α3β2 re-

ceptors was -26.95 ± 1.96 mV (n = 5), for α3β2γ2 receptors it was -20.57 ± 0.77 mV

(n = 4) and for α3β2ε receptors -24.53 ± 6.99 mV (n = 6). These values are consistent with

the chloride equilibrium potential in oocytes (approximately -25 mV) and, therefore, provide

confirmation for chloride-ion flux across the membrane upon agonist binding (Dascal, 1987).

-80 -60 -40 -20 0

-125

-100

-75

-50

-25

0

25

50

α3β2

α3β2γ2α3β2ε

Holding potential (mV)

Nor

mal

ised

cur

rent

Figure 3.6: Current-voltage relationships of GABA-induced currents in Xenopus oocytes ex-pressing α3β2, α3β2ε and α3β2γ2 receptors.Reversal potentials for α3β2, α3β2ε and α3β2γ2 receptors were -26.95 ± 1.96 mV (n = 5),-24.53 ± 6.99 mV (n = 6) and -20.57 ± 0.77 mV (n = 4), respectively. Data represent x-axis in-tercepts ± SEM; n represents the number of oocytes examined. Data were fitted by linear regressionanalysis.

3.2.3 Effect of zinc

It is known that zinc ions inhibit GABAA receptor function and are most potent on α- and

β-subunit-containing receptors and least potent on receptors containing α, β and γ subunits

(Draguhn et al., 1990). Receptors containing the ε or δ subunits have been shown to exhi-

bit medium sensitivity to zinc inhibition (Whiting et al., 1997; Krishek et al., 1998). Thus,

zinc can be used as a marker for subunit co-assembly (i.e. relative insensitivity to zinc in-

68


dicates the existence of the γ subunit in the receptor complex). To verify correct subunit

co-assembling, increasing concentrations of zinc with submaximal GABA concentrations

(EC20) were applied to Xenopus oocytes expressing α3β2, α3β2ε or α3β2γ2 receptor sub-

types. GABA-induced currents in the receptor subtypes examined were inhibited with dis-

tinct potency by zinc (Figure 3.7). GABAA receptors containing the α3β2 subunits were

most sensitive to zinc inhibition with an IC50 value of 0.51 ± 1.18 µM (n = 7). Oocytes

expressing α3β2ε receptors were about 90 times less sensitive to the action of zinc and ex-

hibited an IC50 of 46.67 ± 1.08 µM (n = 8; p < 0.01, unpaired t-test). Receptors containing

the α3β2γ2 subunits were least sensitive to zinc. Their IC50 was 587.49 ± 1.17 µM (n = 7;

p < 0.01, compared with α3β2, unpaired t-test).

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1

0

25

50

75

100


log [Zinc, M]

GAB

A cu

rren

t (%

con

trol

)

Figure 3.7: Zinc concentration-response curves for GABAA receptors expressed in Xenopuslaevis oocytes.Concentration-response curves for the inhibition of GABA (EC20 concentration) by zinc in oocytesexpressing α3β2 (n = 7), α3β2ε (n = 8) and α3β2γ2 receptor subtypes (n = 7). Currents werenormalised to the maximal response induced by GABA. IC50 values are specified in the text (Section3.2.3). Data correspond to means ± SEM; n represents the number of oocytes examined. Curveswere fitted using a non-linear regression fit with variable slopes as described in Section 2.3.5.

3.2.4 Application of the benzodiazepine flunitrazepam

In order to gain further confirmation for the correct subunit co-assembling, modulation by the

benzodiazepine flunitrazepam was tested in oocytes expressing α3β2γ2 and α3β2ε receptor

subtypes, respectively. It is known that the benzodiazepine binding site consists of residues

69


from both α and γ subunits (Pritchett et al., 1989b) and ε-subunit-containing receptors are

insensitive to benzodiazepines (Whiting et al., 1997). Currents evoked by GABA (EC20)

were enhanced with 1 µM flunitrazepam by 68.5 ± 10.01 % (compared to GABA alone)

in oocytes expressing α3β2γ2 receptors (n = 4). This positive receptor modulation further

verified presence and assembly of the γ subunit into the receptor complex. In contrast,

α3β2ε receptors showed no response to 1 µM flunitrazepam administration (97.26± 2.33 %

of GABA control response, n = 6, p < 0.0001; see Figure 3.8).

α3β2γ2 α3β2ε

0

25

50

75

100

125

150

175

200***________________ ___________

% o

f GAB

A co

ntro

l res

pons

e

Figure 3.8: Effect of flunitrazepam on GABAA receptors expressed in Xenopus laevis oo-cytes.Oocytes expressing α3β2γ2 receptors were positively modulated by 1 µM flunitrazepam(168.5 ± 10.01 % of GABA control response, n = 4); data were determined by application of GABA(EC20) alone (control) or co-administered with 1 µM flunitrazepam. In contrast, flunitrazepam didnot affect GABA-control currents in oocytes expressing the α3β2ε receptor subtype (97.26± 2.33 %of control response, n = 6). Currents were normalised to the maximal response induced by GABA atEC20 concentrations. Columns represent mean values (± SEM). *** indicates a significant differencebetween the two receptor subtypes (p < 0.0001).

3.2.5 Constitutive activity of the α3β2ε GABAA receptor subtype

Previous studies have shown that ε-subunit-containing receptors exhibit spontaneous acti-

vity in the absence of GABA (Neelands et al., 1999; Davies et al., 2001; Maksay et al.,

2003; Wagner et al., 2005). It was also noticed that those receptors required an unusually

large holding current to clamp the cells at a specific voltage. The latter was also evident in

70


this study and it was tested whether α3β2ε receptors also exhibit constititive receptor ac-

tivity. Administration of 100 µM picrotoxin to oocytes expressing α3β2ε receptors caused

clear “outward” currents of 21.49± 1.98 % (of total response capacity (maximal GABA res-

ponse + response to picrotoxin = 100 %), n = 8; Figure 3.9 B), indicating agonist-independent

receptor activity. In contrast, picrotoxin had only a marginal effect on oocytes expressing α3

and β2 subunits (n = 5; Figure 3.9 A). Furthermore, it was evident throughout all experi-

ments that oocytes expressing α3β2ε receptors exhibited large baseline (leak) currents when

clamped at -60 mV.

GABA PTX

200 nA25 sec

100 %

21.49 ± 1.98 %

GABA PTX

500 nA

25 sec

A

B

Figure 3.9: Representative current traces evoked by 1 µM GABA and 100 µM picrotoxinin oocytes expressing α3β2 subunits (A) and by 1 mM GABA and 100 µM picrotoxin inoocytes expressing α3β2ε receptor subtypes (B).

3.2.6 Effect of etomidate on α3β2ε GABAA receptors

Anaesthetics, such as etomidate, exhibit potentiation and at high concentrations direct acti-

vation of GABAA receptors (Thompson et al., 1996; Pistis et al., 1997). The receptor mo-

dulation has been shown to be dependent on the type of β subunit expressed in the receptor

71


subtype (Hill-Venning et al., 1997). Here, the influence of the presence of the ε subunit in

the α3β2 receptor complex was investigated in view of modulation and direct activation by

etomidate. Data were compared with those obtained for the α3β2 receptor subtype. Eto-

midate enhanced GABA-elicited currents (EC20) in oocytes expressing the latter by up to

402.1 ± 77.27 % (n = 8; Figure 3.10 A and C). In contrast, the level of etomidate efficacy

was greater in oocytes expressing the α3β2ε receptors (up to 917.3± 83.45 %; n = 7; Figure

3.10 B and C). However, α3β2ε receptors were significantly less sensitive to etomidate (EC50

13.34 ± 1.18 µM; p < 0.0001) compared to α3β2 receptors (EC50 1.71 ± 0.31 µM). Both

receptor subtypes were also activated in the absence of GABA by 100 µM etomidate alone.

Oocytes expressing the α3β2ε receptor showed a greater level of activation (651.5± 57.35 %

of GABA control response (EC20), n = 6) compared to α3β2 receptors (146.7 ± 1.24 % of

GABA control response, n = 3; Figure 3.10 D).

72


α3β2ε α3β20

200

400

600

800 ***______________

% o

f con

trol

GAB

A re

spon

se

-8 -7 -6 -5 -4 -30

200

400

600

800

1000α3β2εα3β2

100

log [Etomidate, M]

% G

ABA

cont

rol r

espo

nse

1000 nA

200 sec

α3β2

GABAEtomidate 0.1 1 10 30 100

1000 nA

200 sec

α3β2ε

GABAEtomidate 1 3 10 30 100

C DA B

C D

Figure 3.10: Effect of etomidate on oocytes expressing α3β2 and α3β2ε receptor subtypes,respectively.(A) and (B) Current traces evoked by application of GABA (EC20) alone or with successive concen-trations of etomidate (in µM) in oocytes expressing the α3β2 (A) and α3β2ε receptor subtypes (B).Horizontal bars represent drug application. (C) Etomidate concentration-response relationships weredetermined by administering increasing concentrations of etomidate with GABA (EC20). Applicationof GABA at EC20 concentrations alone served as control (set to 100 %). Datapoints are means ±SEM (n = 8 for α3β2 and n = 7 for α3β2ε). Curves were fitted using a non-linear regression fitwith variable slopes as described in Section 2.3.5. Currents were normalised to the maximal responseinduced by GABA at EC20 concentrations. EC50 values are specified in the text (Section 3.2.6); nrepresents the number of oocytes examined. (D) Columns represent mean direct activation of the re-ceptors by 100 µM etomidate (± SEM; α3β2 receptor activation: 146.7 ± 1.24 % of GABA controlresponse, n = 3; α3β2ε receptor activation: 651.5 ± 57.35 % of GABA control response, n = 6).Currents were normalised to the maximal response induced by GABA at EC20 concentrations. ***indicates a significant difference between the two receptor subtypes (p < 0.001).

73


3.2.7 Expression of a truncated ε splice variant

Recently, Kasparov et al. (2001) described a truncated ε subunit version (εT) expressed in the

rat brain that exhibits a 26 amino acid deletion starting in the middle of TM2 of the subunit.

The εT subunit was generated using site-directed mutagenesis and co-expressed with α3 and

β2 subunits in Xenopus laevis oocytes, in order to investigate, whether the εT subunit could

form functional channels with α3 and β2 subunits in vitro. Initially, GABA concentration-

response relations were investigated (see Figure 3.11 A). EC50 values for α3β2εT receptors

were 6.48± 0.98 µM (Hill slope: 0.7± 0.24, n = 6). These results were very similar to results

obtained for the α3β2 receptor (EC50: 3.30 ± 1.17 µM, p = 0.07; see Section 3.2.1). Fur-

thermore, the concentration-response curves to zinc inhibition for these two receptor types

were also greatly similar (IC50 0.78 ± 0.84 µM, n= 7, p = 0.33, compared to results obtained

for α3β2 receptors; Figure 3.11 B). These findings suggest that the εT subunit examined

here was probably not incorporated into functional channels in Xenopus laevis oocytes and

channels expressed consisted solely of α3β2 receptors.

-9 -8 -7 -6 -5 -4 -3 -2

0

25

50

75

100

α3β2εT

α3β2εα3β2γ2

α3β2

log [GABA, M]

% m

axim

um r

espo

nse

-9 -8 -7 -6 -5 -4 -3 -2 -1

0

25

50

75

100 α3β2α3β2εα3β2γ2α3β2εT

log [Zinc, M]

% G

ABA

cont

rol r

espo

nse

A B

Figure 3.11: Concentration-response curves for GABA activation and zinc inhibition in oo-cytes expressing α3β2εT receptor subtypes.(A) GABA concentration-response curves of α3β2 (n = 11), α3β2ε (n = 12), α3β2γ2 (n = 10) andα3β2εT receptor subtypes (n = 6). EC50 values are specified in the text (Sections 3.2.1 and 3.2.7).(B) Concentration-response curves for the inhibition of an EC20 concentration of GABA by zinc inoocytes expressing α3β2 (n = 7), α3β2ε (n = 8), α3β2γ2 (n = 7) and α3β2εT receptor subtypes(n = 7). Currents were normalised to the maximal response induced by GABA at EC20 concentra-tions; IC50 values are specified in the text (Sections 3.2.3 and 3.2.7). Data correspond to means ±SEM; n represents the number of oocytes examined. Curves were fitted using a non-linear regressionfit with variable slopes as described in Section 2.3.5.

74


3.3 Discussion

3.3.1 Pharmacological characteristics of the ε subunit

The GABAA receptor ε subunit represents a minor subunit population with restricted ex-

pression pattern in the brain. There have been only a few studies investigating the func-

tions and characteristics of recombinant and native receptors, and the exact role, receptor

subtype structure, and assembling characteristics remain widely unknown. In this study,

co-expression of the ε subunit with the α3 subunit (both genes are clustered on the Xq28

chromosome; Wilke et al., 1997) and the β2 subunit, which is the most widespread β subunit

in the brain (Olsen & Sieghart, 2008), were examined. It is likely that α3 and β2 subunits

form functional receptors with the ε subunit in some areas of the brain. The data obtai-

ned emphasise that the inclusion of the ε subunit into the α3β2 GABAA receptor complex

confers distinct pharmacological properties to the receptor which also differ significantly

from properties of α3β2γ2 receptors.

The inclusion of the ε subunit conferred high GABA sensitivity to the α3β2 receptor subtype

and results obtained were consistent with results seen in previous studies by Neelands et al.

(1999) and Whiting et al. (1997), although in these latter studies the α1 instead of the α3

subunit was expressed with the ε and β3 or β1 subunits, respectively. In contrast, Ranna

et al. (2006) expressed α3β1ε receptor subtypes and reported a reduced sensitivity to GABA

compared to α1- and ε-subunit-containing receptors (Whiting et al., 1997; Neelands et al.,

1999). A study by Ebert et al. (1994) demonstrated reduced GABA sensitivity in α3β1γ2

receptors, compared to α3β2γ2 receptors and it was shown that both α and β subunits are

able to contribute to the affinity to GABA (Ebert et al., 1994; Ducic et al., 1995). This

could explain the differences observed in this study compared to the work by Ranna et al.

(2006). Muscimol is a structural GABA analogue found in the mushroom Amanita muscaria

(Snodgrass, 1978) and its potency is known to be higher when compared to GABA (Harris

et al., 1997). Consistent with this, in this study, all receptor subtypes expressed showed

higher sensitivity to muscimol compared to GABA. Again, ε-containing receptors were the

most potent subtype to the agonist.

75


The Hill coefficients of GABA and muscimol agonist curves measured in oocytes expressing

α3β2ε receptors were lower when compared to Hill slopes of α3β2 and α3β2γ2 receptors.

This was also reported previously by Maksay et al. (2003) and Whiting et al. (1997). The rea-

son for this is not known. In contrast to αβ or αβγ receptors, which usually need binding of

two agonist molecules to sufficiently open the receptor channel, ε-subunit-containing recep-

tors may have only one agonist binding site, as described above (Section 1.2.5.1, Chapter 1;

Jones & Henderson, 2007; Bollan et al., 2008). The low Hill coefficient seen in ε-containing

receptors would support the hypothesis of one agonist binding site in ε-containing receptors,

because the Hill slope reflects the minimal number of interacting binding sites in positively

cooperating reactions (Abeliovich, 2005). However, an additional explanation could be that

the lower Hill slopes are partly due to the increased desensitisation rate seen in ε-subunit-

containing receptors, since desensitisation also influences the Hill coefficient (Celentano &

Wong, 1994; Dominguez-Perrot et al., 1996; Whiting et al., 1997).

Zinc inhibition has been suggested to be a good marker for correct GABAA receptor assem-

bly, as GABAA receptors containing α and β subunits are highly sensitive to zinc inhibition,

whereas receptors which include the γ subunit are less affected by the action of zinc (Dra-

guhn et al., 1990). Receptors containing the δ or ε subunit have been shown to exhibit

medium sensitivity to zinc inhibition. The results obtained here were in consent with this in-

formation. IC50 values for all three receptor subtypes were similar to the results obtained by

Neelands et al. (1999) and Whiting et al. (1997) and indicated correct subunit co-assembling

of the three receptor subtypes by showing significant differences in sensitivity to zinc inhi-

bition.

It is well known that the presence of the γ subunit within the GABAA receptor complex is

needed for benzodiazepine binding to the receptor and recombinant receptors, substituting

the γ with the ε subunit, have been shown to be insensitive to benzodiazepine modulation

(Davies et al., 1997; Whiting et al., 1997). Here, the benzodiazepine flunitrazepam was ap-

plied to further test the correct assembly of α3β2ε and α3β2γ2 receptors. As expected,

GABA-induced currents in α3β2γ2 receptors were potentiated by 1 µM flunitrazepam, whe-

reas the current measured in oocytes expressing the α3β2ε receptor remained unchanged to

76


the GABA-evoked control response. Although there is a lack of a benzodiazepine binding

site in ε-containing receptors, there may be a similar modulatory site for a yet unknown

compound at the interface of ε and α subunits.

Sensitivity of GABAA receptors to the intravenous anaesthetic etomidate is known to be de-

pendent on the β subunit expressed in the receptor subtype, with β2 and β3 subunits being

highly sensitive to etomidate potentiation compared to the β1 subunit, which is least sensi-

tive (Belelli et al., 1997; Hill-Venning et al., 1997). Etomidate potentiates GABAA receptors

at lower concentrations as well as directly activates receptors at high concentrations. It was

shown that the ε subunit confers distinct pharmacological properties in response to etomi-

date when incorporated into the α3β2 GABAA receptor. Currents evoked by submaximal

GABA concentrations (EC20) were enhanced to a greater level in α3β2ε receptors compa-

red to α3β2 receptors. However, inclusion of the ε subunit into the α3β2 receptor complex

resulted in an approximately eightfold higher EC50 value compared to the α3β2 receptor.

Previous studies had also shown that including a third subunit, γ or ε, into the αβ receptor

complex reduces the potency of certain anaesthetic agents (Thompson et al., 1998). Etomi-

date at high concentrations triggered a GABA-mimetic effect in both receptor subtypes. The

potentiation of the GABA-control response, however, was greater in α3β2ε receptors com-

pared to α3β2 receptors. Previous studies by Thompson et al. (1998) demonstrated a greater

level of potentiation in α1β2 compared to α1β2ε receptors. The reason for this difference is

not clear, since the β subunit is the major determinant of sensitivity to the modulating effects

of etomidate (Hill-Venning et al., 1997). Nevertheless, Hill-Venning et al. (1997) showed

that α3-subunit-containing receptors can contribute to differences in maximal potentiation

by etomidate compared to other α subunits, and Ranna et al. (2006) and Thompson et al.

(1998) demonstrated that the inclusion of the ε subunit into αβ1 receptors can have some in-

fluencing effects on GABA-mimetic properties. It was shown that direct receptor activation

by etomidate is normally minimal or absent in receptors containing the β1 subunit. However,

inclusion of the ε into the receptor complex resulted in small direct effects.

As also seen in previous recombinant expression studies, receptors containing the ε subunit

exhibited spontaneous receptor activity (Whiting et al., 1997; Neelands et al., 1999; Maksay

77


et al., 2003). The level of spontaneous activity in α3β2ε receptors was similar to the results

obtained by Mortensen et al. (2003) expressing α1β3ε receptors in Xenopus oocytes. Fur-

thermore, a large holding current was required to sustain a potential of -60 mV in Xenopus

laevis oocytes, which was not seen in the other receptor combinations investigated, indica-

tive for constitutive receptor activity (Neelands et al., 1999; Maksay et al., 2003). This was

confirmed when applying picrotoxin, a channel blocking agent, to oocytes expressing the re-

ceptor, resulting in “outward” currents, which demonstrated the blockade of spontaneously

open channels (Maksay et al., 2003). Several amino acids located in the channel-lining TM2

are involved in chloride conduction, sensitivity to channel blockers and channel gating of

GABAA receptors and the related nAChRs (Filatov & White, 1995; Labarca et al., 1995;

Scheller & Forman, 2002). Different mutations of amino acids in the TM2 have identified

several residues that increase unliganded openings of the receptors and play a role in channel

gating, for example the 9’ leucine residue. Another residue critical for channel gating is the

15’ serine residue. Mutations of the latter amino acid residue have been shown to influence

channel gating and spontaneous receptor activity in β1, α1 and α2 GABAA receptor subunits

(Findlay et al., 2001; Scheller & Forman, 2002; Miko et al., 2004). The ε subunit is the only

GABAA receptor subunit with a glycine residue at this position. Point mutations of the 15’

serine in β1 subunits to glycine or tryptophan increased spontaneous currents blocked by

picrotoxin (Findlay et al., 2001; Miko et al., 2004). Therefore, it is likely that the presence

of the glycine residue at this position entails constitutive receptor activity in ε-containing

receptors (Wagner et al., 2005). As mentioned before, whether or not native ε-containing

receptors exhibit constitutive activity is still not clear, and functional implications of such a

receptor type in vivo would depend on many factors, such as the location of the receptor.

It is important to verify the correct receptor assembly, because it is known that α3β2γ2

receptors can sometimes assemble as a mixture of α3β2 and α3β2γ2 receptors (Boileau

et al., 2002, 2003). However, Jones & Henderson (2007) showed by applying zinc with

high GABA concentrations that the majority of receptors formed by α, β and ε subunits

consisted of all three subunits at a transfection ratio of 1:1:1 and αβ receptors are mostly

absent. To verify correct subunit co-assembling, it was shown that α3β2γ2 receptors were

78


relatively insensitive to zinc and GABA-control responses were enhanced by application of

flunitrazepam. α3β2ε receptors were also assembled correctly, which was demonstrated

by medium zinc sensitivity, an outward current after the application of the channel blocker

picrotoxin, a high leak current when voltage clamped at -60 mV and relatively small GABA-

evoked currents compared to currents obtained from other receptor subtypes.

3.3.2 Investigation of a truncated ε subunit variant

Several previous studies reported a number of different splice variants seen in the human and

rodent ε subunit (Whiting et al., 1997; Sinkkonen et al., 2000; Kasparov et al., 2001). In

rodent brain, an appropriate spliced mRNA, a large version, including a Pro/Glx motif, and

a truncated version have been described. The latter exhibits a 26 amino acid deletion in the

second putative TM domain. Since such a deletion would have a major impact on receptor

structure and function, and the role of the ε subunit is still not understood, it was investigated

whether the truncated version would form functional receptors with the α3 and β2 subunit

in Xenopus laevis oocytes. GABA concentration-response relationships were very similar

to the ones seen in α3β2 receptors. The same applied to the inhibition of the receptor by

zinc. These results suggested that the truncated subunit is not incorporated into the receptor,

which does not seem to come as a surprise, due to the fact that amino acids of the TM2

and the entire extracellular loop between TM2 and TM3 are missing. Those amino acids

have been shown to be important for channel pore formation (TM2) and channel gating upon

agonist binding (extracellular loop between TM2 and TM3). Therefore, a deletion would

have a major impact on receptor structure and function (Smith & Olsen, 1995; Kash et al.,

2004). Furthermore, a recently published study by Pape et al. (2009) showed Western blot

analysis of ε subunit expression in the spinal cord and hypothalamus of twelve days old rats.

The results revealed a 70 kDa band (appropriate spliced ε subunit) and a large ε subunit

version of 120 kDa (the large version described by Moragues et al. (2000)), but no band

which would correspond to the truncated version. These findings indicate that the truncated

version is not translated into proteins in these brain regions, which would support our results.

Although a number of ε splice variants have been investigated, none of the splice forms have

79


been shown to be incorporated into functional receptors in vitro or in vivo up until now.

3.3.3 Summary

In summary, the rat GABAA receptor ε subunit as well as a truncated ε splice version together

with the α3 and β2 subunits were co-expressed in Xenopus laevis oocytes and pharmacolo-

gical properties of α3β2ε receptors were examined. This subunit combination has not been

investigated previously, but it is likely that α3β2ε receptors exist in certain areas of the brain

in vivo. Data obtained revealed that the α3β2ε receptor exhibits unusual pharmacological

characteristics, previously described in other ε-subunit-containing receptor subtypes. Fur-

thermore, it was shown that receptors were enhanced to a greater level by the anaesthetic

etomidate compared to α3β2 receptors. However, as stated previously, inclusion of a third

subunit into the αβ receptor complex also lowered etomidate potency in α3β2ε receptors.

So far, no subunit-selective pharmacology could be recognised. Nevertheless, the ε subunit

may represent an important pharmaceutical target for future drug development and further

research is needed to uncover the promiscuous role of this subunit. It was also shown that

the truncated ε splice variant was not included into functional α3β2 receptor channels in

Xenopus laevis oocytes.

80

Chapter 4

Development of an assay to screen

compound libraries

4.1 Introduction

The structural diversity of GABAA receptors and the plethora of drug binding sites make

them attractive targets for drug discovery, especially since it was demonstrated that the va-

rious subunits regulate different drug effects (Section 1.2.4, Chapter 1; Möhler, 2006). Al-

though several studies have been conducted in order to identify the pharmacological profile

of ε-subunit-containing receptors, no subunit-specific pharmacology could be identified (Da-

vies et al., 1997; Whiting et al., 1997; Thompson et al., 1998; Neelands et al., 1999; Maksay

et al., 2003; Ranna et al., 2006). However, these receptors possibly have specific modulatory

binding sites which may be important for future drug development. The next aim of this pro-

ject was to set up a cell-based system to screen the ε-containing receptor against compounds

from a commercially available library and, with luck, to identify compounds that modulate

this specific GABAA receptor subtype.

As described in Section 1.3, Chapter 1, screening ion channels for modulating compounds is

a challenging research area, however, several assay technologies are available. Electrophy-

siological recordings are the gold standard method to measure ion channel function (Smith

& Simpson, 2003; Verkman & Galietta, 2009). However, they are very labour intensive

81

Chapter 4. Assay development

and, hence, time consuming. Therefore, in the majority of studies the primary screening for

compounds is performed via high throughput assays to detect lead compounds. Many scree-

ning methods require specialised measurement devices, such as voltage/ion probe readers,

FLIPRs or scintillation counters and only few can be set up easily and at low costs using

commercially available measuring devices. In this project, a suitable method was sought to

screen ε-subunit-containing GABAA receptors for novel modulatory compounds. Such com-

pounds are thought to be important to uncover the role of the GABAA receptor ε subunit and

they may even act as future drugs on these receptor subtypes. The assay technologies applied

in this study were first set up using α3β2 or α3β2γ2 receptors before testing constitutively

open ε subunit receptors, which may add some additional challenges to the setup.

4.1.1 Clomeleon

GFP was originally isolated from jellyfish (Aequorea victoria) and has been used extensively

in the field of molecular biology as a fluorescence probe (Tsien, 1998; Wachter et al., 1998).

YFP, a derivative of GFP, exhibits enhanced brightness and a longer wavelength compared

to other GFP variants. It is the only fluorescent protein that can be quenched by small anions

and, thus, it can be used as a chloride-ion indicator (Wachter & Remington, 1999; Jayara-

man et al., 2000). Kuner & Augustine (2000) have constructed a FRET-based, ratiometric,

genetically-encoded chloride-ion indicator protein termed Clomeleon. It was generated by

linking a CFP, which is insensitive to chloride, to a YFP using a flexible recombinant To-

bacco Etch Virus (rTEV) protease peptide linker. In the absence or at low concentrations of

chloride in the cell, upon excitation of the CFP, Clomeleon emits energy to the YFP, which

can be measured at ~530 nm (Figure 4.1 A). At rising concentrations of chloride (e.g. upon

GABAA receptor activation), chloride binds to the YFP. This changes its absorption coeffi-

cient and leads to a reduction of the FRET efficacy and increased CFP emission that can be

measured at 485 nm (Figure 4.1 B, C). The ratio of fluorescence emission of the YFP to that

of the CFP is chloride concentration dependent.

Clomeleon has been shown to be a good indicator of intracellular chloride concentrations

in living cells, for example neurons, as well as in transgenic mice, where it was used to

82


440 nm

485 nm 530 nm

CFP YFP

440 nm

530 nm

CFP YFP

485 nm

A B

Figure 4.1: Clomeleon is a FRET-based chloride-indicator protein consisting of two distinctfluorescence proteins.(A) At no or low concentrations of chloride, the excited CFP (donor) transfers most of its energyto the linked YFP (acceptor). (B) If chloride (depicted as orange circle) binds to the YFP, FRETefficiency is reduced leading to increased CFP emission. (C) Emission spectrum of Clomeleon in thepresence of different concentrations of chloride ions (in mM; from Kuner & Augustine, 2000).

83


monitor chloride distribution and concentration changes (Kuner & Augustine, 2000; Haver-

kamp et al., 2005; Duebel et al., 2006; Berglund et al., 2008). It was also applied to measure

chloride shifts after GABAA receptor activation in cultured hippocampal neurons (Kuner &

Augustine, 2000). However, it has not been used in compound screening assays before. Clo-

meleon has several advantages over other chloride indicators: it has a good signal-to-noise

ratio, cannot be influenced by other physiologically important anions, it can be easily loaded

into cells via transfection methods, its fluorescence is very stable and it can be targeted to

specific locations/cells (Kuner & Augustine, 2000; Bregestovski et al., 2009). Its relative

large size (~59 kDa) prevents diffusion from the cell and it is possible to determine abso-

lute chloride concentration changes. However, there are several disadvantages, including its

sensitivity to pH changes, the slow response kinetics and low chloride affinity as well as the

different bleaching of the CFP and YFP. This could lead to an overestimation of the chlo-

ride concentration and, thus, careful setup and optimisation are important for a successful

application.

In this study, it was examined whether Clomeleon can be used in a cell-based screening assay

by expressing it together with GABAA receptors in HEK293 cells. The cells were seeded into

96-well microtitre plates and it was tested whether GABAA receptor activation would result

in a fluorescence shift from 530 to 485 nm by comparing the baseline fluorescence before

adding the receptor agonist with the fluorescence signal afterwards. If GABA application

could elicit a change of the fluorescence signal, the compound screening assay could be

optimised.

4.1.2 Iodide-flux method

Chloride ion channels are not only permeable to chloride, but also to other small anions, na-

mely NO3-, Br-, I-, HCO3

-, SCN- and small organic acids (Tang & Wildey, 2004; Verkman &

Galietta, 2009). However, chloride is the most abundant anion in biological systems, hence,

the name chloride channels. Anions can pass through the channels in either direction (influx

or efflux) according to the electrochemical driving force. Movement of ions can be directly

quantified using ion flux assays, which can be used to characterise ion channel functions.

84


Flux assays often employ radiolabelled tracers, such as 36Cl- or 125I-. Ion channels are acti-

vated in the presence of a known concentration of tracer ions and ion flux through the opened

channels takes place. Subsequently, the medium is moved to a new plate, cells are lysed and

radioactive tracer concentration in both cell lysate and medium, are measured. Receptor ac-

tivity, inhibition or modulation can be detected by determining the ratio of intracellular to

extracellular radioactivity. The signal strength is dependent on tracer ion concentration and

the number of cells (Gill et al., 2003). Screening assays using radiolabelled tracer molecules

are robust, easy to perform and provide data with high sensitivity. However, there are some

major concerns with this method. The application of radioactive tracers raises safety issues

and also implies relatively high costs of the material and the special requirements for its

disposal.

In this project a non-radioactive, colorimetric iodide-flux assay was tested, where iodide acts

as the indicator of chloride channel activity (Tang & Wildey, 2004, 2006). The assay was

derived from the Sandell-Kolthoff (SK) reaction (Sandell & Kolthoff, 1937; Tang & Wildey,

2004):

Ce4+(yellow) + As3+ I-→ Ce3+(colourless) + As4+ (4.1)

The principle of the reaction is that the iodide catalyses the conversion of yellow-coloured

cerium ions (Ce4+) to colourless Ce3+ by arsenic III acids. The remaining Ce4+ in the sample

can be detected by reading the absorbance at a wavelength of 405 nm. The OD405 readings

enable to draw conclusions to the intra- and extracellular iodide concentrations and, thus, ion

flux upon receptor activation (Figure 4.2).

The colorimetric ion flux technology has the advantages of lower costs and improved sa-

fety. Furthermore, the method does not require a confluent cell monolayer and already small

concentrations of iodide can be detected due to the low intracellular concentration (Tang

& Wildey, 2004). Since test compounds are removed from the cells after incubation, there

is also the benefit that they cannot interfere with the assay readout of the lysed cells, a

problem that commonly leads to false positive results in other screening formats, such as

fluorescence-based methods. However, flux assays have certain disadvantages, including

85


Ce4+

Ce3+

(yellow)I-

(clear)

GABAA receptorGABAIodide

A B

Figure 4.2: Assay principle of the colorimetric iodide-flux assay.(A) Iodide-loaded cells in iodide-free solution. (B) Channels open upon agonist binding and iodideexerts into extracellular solution due to electrochemical gradient (adapted from Lynch, 2005)

low temporal resolution, the long cell-loading incubation times, low informational content,

reduced reproducibility and a high signal-to-noise ratio that is influenced by the channel

expression levels (Xu et al., 2001; Gill et al., 2003). A further drawback of the assay is the

relative short stability of the sample-buffer mixture making a fast reading and optimal timing

necessary. The high signal-to-noise ratio is presumably the reason why the assays Z’-factor

obtained by Tang & Wildey (2004) was relatively low (0.42), which indicates a small signal

detection window (Zhang et al., 1999). The Z’-factor provides a mean to assess the quality

and optimisation of an assay in regard to the assay’s signal dynamic range as well as data

variation (Zhang et al., 1999). Z’-factors between 1 and 0.5 represent an assay with excellent

quality, whereas smaller values indicate a smaller signal separation between the positive and

negative controls. Values near zero or negative Z’-factors indicate that the assay was either

not optimised or that a robust setup is not achievable in this format.

In this project, the iodide-flux assay was performed and optimised by expressing α3β2γ2

GABAA receptors in HEK293 cells, which were seeded into 96-well plates. Cells were loa-

ded prior to the assay with iodide. Receptors were activated and iodide efflux from the cells

was determined by performing the SK assay with both, cell lysates and supernatants. By

assessing the OD405 readings, the iodide concentration in the samples, which gives infor-

mation on the receptor activity, could be determined (Tang & Wildey, 2004; Lynch, 2005).

86


4.1.3 YFP-based assays

As displayed below, the attempted assays described before were not suitable for screening

of GABAA receptors transiently expressed in HEK293 cells. Therefore, a further method

was examined applying a YFP-based assay format. This assay was similar to the first as-

say examined, applying the chloride-indicator protein Clomeleon (Section 4.1.1), as it was

based on the halide sensitivity of YFPs. However, the assay used a single fluorophore with

increased sensitivity to halides that could be used to detect anionic influx into a cell using a

fluorescence plate reader (Kruger et al., 2005).

Due to the fact that YFP has a low sensitivity to chloride, YFP variants were constructed,

among others the YFP-H148Q (Wachter & Remington, 1999). This mutant version showed

an improved sensitivity to chloride (Kd 100 mM; Kd represents the anion concentration at

a half-maximal YFP quench) as well as iodide (Kd 21 mM; Jayaraman et al., 2000). Io-

dide does not occur naturally in cells and can be toxic at high concentrations. In order to

prevent such effects, solutions applied to cells should contain low iodide concentrations.

Hence, sensitivities for chloride and iodide of YFP-H148Q were still too low and YFPs with

enhanced properties were sought. Galietta et al. (2001a) generated random mutations of

certain amino acid residues at the halide binding site. This approach identified several mu-

tant YFP variants that showed improved halide sensitivity. The mutant version YFP-H148Q/

I152L exhibited a very high iodide affinity (Kd 3 mM) as well as improved chloride sensiti-

vity (Kd 88 mM). Performing assays with iodide rather than chloride has the advantage that

anion co-transporters or exchangers, for example NKCC (Na+-K+-Cl- co-transporter) or AE1

(anion exchanger 1), transport iodide less efficiently than chloride (Galietta et al., 2001b).

In this project, the YFP mutant version YFP-H148Q/I152L was used as a chloride-indicator

protein in a cell-based screening assay as described by Kruger et al. (2005). It was shown

that YFP-H148Q/I152L is suitable for screening assays by permitting detection of fluores-

cence signal changes upon GABAA receptor activation. This is because, in a medium that

contains a high concentrations of iodide, activation of the GABAA receptor will result in a

reduction of intracellular fluorescence as iodide enters the cell, which can be measured using

87


a fluorescence plate reader (Figure 4.3).

Iodide.GABAA receptor; GABA;/ mutant YFP;

intracellular

extracellularA B

Figure 4.3: Schematic diagram showing the principle of the YFP-H148Q/I152L screeningassay.(A) Cells co-expressing the mutant YFP-H148Q/I152L and GABAA receptors were present in aniodide-containing solution. (B) GABA application triggers the channels to open and iodide entersthe cells upon the electrochemical gradient. Subsequently, the YFP-H148Q/I152L inside the cells isquenched, reducing the fluorescence signal emitted.

This assay format has a range of advantages over other existing methods: (1) it does not

require time-consuming cell loading or washing steps, (2) the YFP mutant version exhibits

excellent optical properties and photostability, (3) it can be excited in the visible range of

light, (4) it is genetically encoded and can be targeted to the required cell type, (5) its kinetics

are faster compared to voltage-sensitive dyes, (6) due to its high molecular weight of 27 kDa,

it is retained in cells and (7) stable expression of the protein(s) of interest is not required

(Galietta et al., 2001b; Bregestovski et al., 2009; Gilbert et al., 2009a). A disadvantage of the

YFP-H148Q/I152L is its relatively slow kinetics (Bregestovski et al., 2009). Nevertheless,

different YFP mutant variants have successfully been used in screening assays of cystic

fibrosis transmembrane conductance regulator protein (CFTR) chloride transport agonists

(Galietta et al., 2001a). Furthermore, it was shown that the mutant YFP is suitable for HTS

of GlyR and GABAA receptors (Kruger et al., 2005; Gilbert et al., 2009b).

88


4.2 Results

4.2.1 Evaluating a screening assay using Clomeleon

The chloride-sensitive protein Clomeleon was first scanned for its excitation and emission

spectrum using a spectrofluorometer. This was to test protein functionality, because the

DNA sequence acquired from sequencing differed in one amino acid from the sequence

received from Dr Thomas Kuner. HEK293 cells were transfected with Clomeleon alone or

Clomeleon and α3β2 subunit receptors; untransfected cells served as negative control. Cells

were excited at a wavelength of 440 nm and emission was scanned over a range of 460

to 650 nm. The obtained spectrum showed a large peak at a wavelength of 527 nm and a

smaller peak at 475 nm (Figure 4.4). Control scans with untransfected cells, mammalian

Ringer solution and the cuvette alone resulted in one or two minor peaks in case of the first

two controls, which were of different wavelengths compared to the two large peaks obtained

for Clomeleon (Figure 4.4). An excitation spectrum obtained for wavelengths from 350 to

450 nm, when measuring emission at 527 nm, showed no clear peak, but a gradual increase

of the signal (Figure 4.5). Furthermore, it was tested whether a shift in the fluorescence

signal could be measured by manually adding GABA (100 µM final concentration) to the

cuvette, but no change of the Clomeleon emission could be detected. It was suggested that

receptor activity may be too fast paced to see a change using the equipment.

Excitation and emission scans of Clomeleon were similar to those described by Kuner & Au-

gustine (2000) and the protein was, therefore, further tested in a microtitre screening assay

setup using a fluorescence plate reader. Clomeleon was expressed in HEK293 cells together

with GABAA receptor α3 and β2 subunits. On the day of the assay, baseline fluorescence

was measured for 20 seconds, before adding 10 µl mammalian Ringer solution with GABA

(100 µM final concentration) to each well and further reading the fluorescence signal for

another 30 seconds at either 527 or 480 nm. Readings obtained from the assay setup did not

show the expected fluorescence shift from 527 nm to 480 nm. Instead, there was little or no

change at both wavelengths. Figure 4.6 A represents the normalised gross fluorescence si-

gnals obtained at 527 as well as 480 nm over 30 seconds after addition of the agonist GABA.

89


450 475 500 525 550 575 600 625 650

0

50000

100000

150000

untransfected cells

Clomeleon

Wavelengh (nm)

Counts [1/sec]

450 475 500 525 550 575 600 625 650

0

25000

50000

75000

100000

125000

Clomeleon & α3β2

mammalian Ringer

cuvette

Wavelengh (nm)

Counts [1/sec]

A

B

Figure 4.4: Emission scan of the chloride-indicator Clomeleon expressed in HEK293 cells.Fluorescence signal of cells expressing Clomeleon versus untransfected cells (A) and spectrum ofcells expressing Clomeleon and α3β2 GABAA receptors as well as of mammalian Ringer solutionalone and a cuvette only (B). Dashed lines highlight the two peaks of the fluorescence signal emittedby Clomeleon.

90


350 375 400 425 450 475

0

5000

10000

15000

20000

25000

Clomeleon & α3β2

Wavelength (nm)

Counts [1/sec]

Figure 4.5: Excitation spectrum of HEK293 cells expressing Clomeleon and α3β2 GABAA

receptors.

Figure 4.6 B displays the normalised net fluorescence signals after subtracting the signal of

untransfected control cells. Data depicted show a small, non-significant decrease of fluores-

cence signal compared to baseline fluorescence after agonist injection at both wavelengths.

This was most likely an artefact of the assay setup, since the signal at both emission wave-

lengths decreased and no increase could be seen for CFP emission. Altering cell densities

could not improve the measurement’s outcome.

91


60

65

70

75

80

85

90

95

100

105

5 10 15 20 25 30

480 nm (n = 4)527 nm (n = 10)

BF

Time [sec]

% g

ross

flu

ore

sce

nce

A

60

65

70

75

80

85

90

95

100

105

5 10 15 20 25 30

527 nm (n = 6)480 nm (n = 4)

BF

Time [sec]

% n

et

flu

ore

sce

nce

B

Figure 4.6: Fluorescence signal of Clomeleon in HEK293 cells also expressing α3β2GABAA receptors after addition of GABA.(A) Gross fluorescence signal. (B) Net fluorescence signal. Fluorescence was measured at emissionwavelengths of 527 nm or 480 nm for 30 seconds after injecting GABA (100 µM final concentration).Data correspond to means ± SEM; n states the number of wells measured. Fluorescence was nor-malised by setting the averaged baseline fluorescence (BF) to 100 % and fluorescence changes (afteradding GABA) were calculated as percentage of the baseline fluorescence.

92


4.2.2 Iodide-flux assay

Due to the fact that the Clomeleon-based assay setup did not provide the required outcome, a

colorimetric iodide-flux screening assay, based on the modified SK method, was tested next.

Tang & Wildey (2004) demonstrated that a modified SK reaction could indicate GABAA

receptor activity in WSS-1 cells by measuring iodide concentration changes after incuba-

tion of iodide-loaded cells with the agonist GABA. In order to determine the assays linear

range and sensitivity, standard curves were established by measuring the OD405 of increa-

sing concentrations of NaI subjected to the SK reaction (Figure 4.7). The linear range of the

assay was estimated from the standard curves in Figure 4.7 B; it was between 300 nM and

3 µM after 15 minutes incubation time and between 100 nM and 1 µM after 30 minutes incu-

bation time, demonstrating a shift towards the left, depending on the assay incubation time.

Further, it was evident that the OD405 readings of the SK reaction decrease with incubation

time (Figure 4.7 B). The assays lowest detection limit after 15 minutes incubation time was

estimated from the standard curve data as described in Section 2.6.2. The OD405control was

2.972 with a SD of 0.030. The determined lowest detection limit was 407 nM (OD405 at

407 nM was 2.88, ∆OD405 was 0.092 and 3SDControl was 0.09; Figure 4.7 A).

After obtaining a standard curve, HEK293 cells were transfected with α3β2γ2 GABAA re-

ceptors and seeded into 96-well microtitre plates. At first it was tested whether it was pos-

sible to measure a GABA concentration-dependent effect performing the assay as described

by Tang & Wildey (2004, 2006). Figure 4.8 and Table 4.1 illustrate the results obtained from

cell lysates subjected to the SK assay. Those did not show a clear GABA concentration-

dependent pattern, but rather random values. The same applied to the % activity of the

receptors which was calculated as described in Section 2.6.2 (Tang & Wildey, 2004). Fur-

thermore, OD405 values of untransfected control cells were expected to be low, due to a high

iodide concentration in the cells, however, the values obtained were relatively high and si-

milar to the data from cells expressing α3β2γ2 receptors. It is worth mentioning that results

shown were not reproducible, but varied widely, despite repeating the assay several times.

The OD405 of the cell supernatant, the DPBS solution containing different agonist concen-

93


-9 -8 -7 -6 -5 -4

0

1

2

3 10 min

15 min

20 min

25 min

30 min

35 min

40 min

70 min

log [NaI, M]

OD 405

-7.5 -7.0 -6.5 -6.0 -5.5

0

1

2

3

log [NaI, M]

OD 405

A

B

Figure 4.7: NaI flux assay standard curve. (A) NaI concentrations of 3 µM to 25 nM wereapplied to the modified SK method. OD405 was read after 15 minutes incubation time. (B)The modified SK assay was performed on samples containing different NaI concentrations(300 µM to 3 nM). OD405 was determined after different incubation times ranging from 10to 70 minutes. Data were averaged from eight data points ± SEM.

94


0

100 pM

1 nM

10 nM

100 nM Mµµµµ

1 Mµµµµ

10

Mµµµµ

100 1

mM

10 mM

10 mM 0

0

1

2

3

4

*

*

[GABA]

OD405

Figure 4.8: A modified SK assay was used to measure chloride conductivity in HEK293cells expressing α3β2γ2 GABAA receptors (yellow columns) and untransfected HEK293cells (white columns).Before the assay, cells were loaded with iodide-loading buffer and incubated with increasing concen-trations of GABA as described in the text. OD405 was measured after 15 minutes incubation of SKreaction. Values were means ± SD from 16 data points from two different microtitre plates. * indi-cates p < 0.05 compared to highest GABA concentration of 10 mM.

Table 4.1: OD405 readings and analysis of results obtained from modified SK assay inHEK293 cells expressing α3β2γ2 GABAA receptors.

α3β2γ2 receptors untransfected cellsGABA 0 µM 100 pM 1 nM 10 nM 100 nM 1 µM 10 µM 100 µM 1 mM 10 mM 10 mM 0 µM

OD405 mean 2.387 1.854 2.279 2.303 1.696 1.740 1.904 2.037 1.927 1.857 1.321 1.706SD 0.464 0.936 0.579 0.575 0.917 0.747 0.907 0.957 0.859 0.691 0.636 1.029

% activity 0.00 100.66 20.28 15.73 130.41 122.06 91.12 65.94 86.84 100.00

Cells were loaded with iodide-loading buffer and incubated with increasing concentrations of GABAas described in the text. The SK assay was conducted for 15 minutes before determining OD405readings. Mean values were calculated from 16 data points from two different microtitre plates, %activity as described in Section 2.6.2.

95


trations added to the iodide-loaded cells, was also investigated. After the incubation time,

the supernatant was transferred to a separate plate and further subjected to the SK assay. De-

pending on GABAA receptor activity, the supernatant should contain distinct concentrations

of iodide. Compared to the results obtained from lysed cells, the OD405 readings of the

supernatant were very low throughout, indicating a high extracellular concentration of io-

dide (Figure 4.9 A). Because the assay has a relatively small detection range (see above) and

iodide concentration of supernatant seemed to be very high in all samples, it was examined

whether diluting the supernatants would improve the results. It was therefore transferred to a

new plate after incubation and diluted 1:2 or 1:5. Although the OD405 readings were higher

with increased dilution, the results did again show no concentration-dependent pattern but

seemed arbitrary (Figure 4.9 B, C). Furthermore, supernatants from untransfected control

cells had a relatively low OD, indicating iodide efflux from the cells into the supernatant via

other means than the examined GABAA receptors.

GABA incubation time for iodide-loaded cells was examined by performing a time-response

course in order to establish the optimal time to improve the assay outcome. After washing

the cells, they were incubated with 10 µM GABA in DPBS or DPBS only for 10, 20, 30 or

45 seconds or 1, 2, 3 or 5 minutes. The results obtained from the cell lysate, after performing

the SK assay, are shown in Figure 4.10. OD405 readings for cells treated with GABA or

DPBS were similar and overall relatively high. Figure 4.11 represents OD405 values from

the supernatants (undiluted or diluted) after performing the SK assay. Samples from the

undiluted plate have low OD405 values throughout, indicating a high iodide concentration.

Therefore, samples were diluted in the following tests to be able to distinguish between

different iodide concentrations. Although there were significant differences between GABA

and DPBS treated cells in some cases, results obtained varied on each plate tested (see also

Figures B.3 and B.4, Appendix B). Due to the discrepancies obtained throughout, it was

difficult to decide on an optimal GABA incubation time. Therefore, the incubation time of

five minutes applied by Tang & Wildey (2004) was further applied.

96


0

100 p

M1 n

M10

nM

100 n

M Mµµµµ1

Mµµµµ10

Mµµµµ

100 1 m

M

10 m

M

10 m

M 00.0

0.2

0.4

0.6

0.8

1.0

[GABA]

OD

405

A

0

100 p

M1 n

M10

nM

100 n

M Mµµµµ1

Mµµµµ10

Mµµµµ

100 1 m

M

10 m

M

10 m

M 00.0

0.5

1.0

1.5

2.0

*

** ***

*

[GABA]

OD

405

0

100 p

M1 n

M10

nM

100 n

M Mµµµµ1

Mµµµµ10

Mµµµµ

100 1 m

M

10 m

M

10 m

M 00.0

0.5

1.0

1.5

2.0

2.5

***

**

[GABA]

OD

405

B

C

Figure 4.9: SK assay determining iodide efflux from HEK293 cells expressing α3β2γ2GABAA receptors (yellow columns) and untransfected HEK293 cells (white columns).Before the assay, cells were loaded with iodide-loading buffer and incubated with increasing concen-trations of GABA as described. The solution containing GABA was transferred to a separate plate.The SK assay was performed on undiluted (A), 1:2 (B) and 1:5 (C) diluted supernatant. After 15minutes incubation the OD405 was measured. All values were means± SD from eight data points onone microtitre plate. * indicates p < 0.05, ** = p < 0.01, *** = p < 0.001 compared to highest GABAconcentration of 10 mM.

97


10 sec

20 sec

30 sec

45 sec

1 min

2 min

3 min

5 min

10 sec

20 sec

30 sec

45 sec

1 min

2 min

3 min

5 min

0

1

2

3

4

10 µM GABA DPBS

*

Time

OD405

Figure 4.10: OD405 values from iodide-flux assay in HEK293 cells expressing α3β2γ2GABAA receptors. Before performing the SK assay, 10 µM GABA (yellow columns) orDPBS (white columns) were incubated for ten seconds to five minutes, respectively.Readings represent data from SK assay from undiluted cell lysate. Values are means ± SD from sixdata points each on three microtitre plates. * indicates p < 0.05 comparing GABA-treated cells withtheir respective DPBS controls.

98

Chapter4.A

ssaydevelopm

ent

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

0.0

0.2

0.4

0.6

0.8

1.010 µM GABA DPBS

***

Time

OD

405

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

0.0

0.5

1.0

1.510 µM GABA DPBS

*****

Time

OD

405

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

0

1

2

3

410 µM GABA DPBS

******

*

TimeO

D40

5 *

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

10 s

ec20

sec

30 s

ec45

sec

1 m

in2

min

3 m

in5

min

0

1

2

3

410 µM GABA DPBS

***** * ***

Time

OD

405

A C

B D

Figure 4.11: Iodide-flux assay. HEK293 cells expressing α3β2γ2 GABAA receptors were exposed to GABA for different incubation times.Before performing the SK assay, 10 µM GABA in DPBS (yellow columns) or DPBS only (white columns) were incubated for ten seconds to five minutes. DPBSwas transferred to new plates and the SK assay was performed on undiluted supernatant (A), on supernatant diluted 2:3 (B), 1:2 (C) or 1:5 (D). OD405 valuesare means ± SD from six data points. * indicates p < 0.05, ** = p < 0.01, *** = p < 0.001 comparing GABA-treated cells with their respective DPBS-incubatedcontrols.

99


In a further attempt, the influence of the cell density was examined. The effect of different

concentrations of GABA at a cell density of 25 × 103 cells per well was compared with its

effect on decreased densities of 17.5 × 103 and 12.5 × 103 cells per well. The SK assay was

performed on lysed cells as well as the undiluted supernatant. The latter showed very low

OD405 readings throughout, indicating a high concentration of iodide (Figure 4.12 B). In

contrast, the readings for the lysed cells differed between different cell densities and GABA

concentrations. In general, the readings had a tendency towards higher values for higher

GABA concentrations, indicating a low iodide concentration inside the cells (Figure 4.12 A),

which would be the desirable assay outcome, but the results obtained were not consistent.

Due to the fact that the data obtained so far did not show well-defined results, it was investi-

gated whether the observed iodide efflux from HEK293 cells was GABA receptor-mediated

by applying bicuculline and picrotoxin to the cells. Bicuculline is a competitive antagonist

of GABAA receptors. Treating the receptors expressed with bicuculline would inhibit the io-

dide flux compared to GABA alone in case of a GABA-mediated response. Exposing cells to

the non-competitive channel blocker picrotoxin would indicate GABAA receptor blockade.

The results showed comparatively high readings for all samples of cell lysate, indicating

low iodide concentration inside the cells (Figure 4.13 A). There were some significant dif-

ferences between the various treatments. However, the values were random, for example

OD405 readings of cells exposed to 1 µM GABA were significantly higher than those of

cells treated with 100 µM, although there should have been lower iodide concentrations in

the latter cells. Readings obtained from the cell supernatant solutions showed low OD405

values for all wells, except for those containing DPBS with GABA and picrotoxin (Figure

4.13 B). The latter exhibited similar OD405 readings to the data obtained for cell lysates,

indicating a low iodide concentration in the cell supernatants, thus blocking iodide efflux

from those cells. Untransfected control cells were treated with GABA alone or GABA and

picrotoxin and showed similar OD405 values as transfected cells. Interestingly, in picrotoxin

treated control cells iodide efflux was also blocked (Figure 4.13 B).

Although optimisation of the iodide-flux assay setup was tested extensively, results varied

and no robust assay setup could be achieved. Since iodide efflux was high in transfected and

100


1 m

M

100

µM

10 µ

M1

µM

1 m

M

100

µM

10 µ

M1

µM

1 m

M

100

µM

10 µ

M1

µM

0.0

0.5

1.0

1.5

2.0

2.512500 cells 17500 cells 25000 cells

***

****

** **

[GABA]

OD

405

1 m

M

100

µM

10 µ

M1

µM

1 m

M

100

µM

10 µ

M1

µM

1 m

M

100

µM

10 µ

M1

µM

0.0

0.2

0.4

0.6

0.8

1.012500 cells 17500 cells 25000 cells

**

[GABA]

OD

405

A

B

Figure 4.12: SK assay measuring iodide flux in HEK293 cells expressing α3β2γ2 GABAAreceptors. Different cell densities were seeded: 12.5 × 103 cells per well (yellow columns),17.5 × 103 cells per well (blue columns) and 25 × 103 cells per well (white columns).Before the assay, cells were loaded with iodide-loading buffer and incubated with increasing concen-trations of GABA as described. The SK assay was performed on undiluted cell lysate (A) and su-pernatant (B). After 15 minutes incubation, the OD405 was measured. All values were means ± SDfrom eight data points on one microtitre plate. * indicates p < 0.05 and ** = p < 0.01 compared tohighest GABA concentration of 1 mM.

101


DPBS

M G

ABA

µµµµ1

M G

ABA+50

µM B

IC

µµµµ1 1 µM

GABA+5

0 µM

PTX

M G

ABA

µµµµ10

0 M

GABA+5

0 µM

BIC

µµµµ10

0

M G

ABA+50

µM P

TX

µµµµ10

0 1 µM

GABA+5

0 µM

PTX

1 µM

GABA

0

1

2

3

4

*** *** ***

*

OD

405

*

A

B

PBS

M G

ABA

µµµµ1

M G

ABA+50

µM B

IC

µµµµ1 1 µM

GABA+5

0 µM

PTX

M G

ABA

µµµµ10

0 M

GABA+5

0 µM

BIC

µµµµ10

0

M G

ABA+50

µM P

TX

µµµµ10

0 1 µM

GABA+5

0 µM

PTX

1 µM

GABA

0

1

2

3 ***

**

***

OD

405

Figure 4.13: OD405 readings of SK assay after applying GABA, bicuculline and picrotoxinto HEK293 cells expressing α3β2γ2 GABAA receptors.Cells were loaded with iodide-loading buffer and subsequently incubated with 50 µM bicuculline or50 µM picrotoxin and 1 µM GABA (yellow columns) or 100 µM GABA (blue columns). Untransfec-ted control cells were also exposed to picrotoxin and 1 µM GABA (white columns). The SK assaywas performed on undiluted cell lysate (A) and cell supernatant (B). After 15 minutes incubation time,OD405 was measured. All values were means ± SD from eight data points each on two microtitreplates. * indicates p < 0.05, ** = p < 0.01 and *** = p < 0.001 compared to DPBS treatment only(grey column). OD405 readings from experiments using PTX were also compared.

102


untransfected cells and the last assay attempt using the channel blocker picrotoxin indicated

the possibility that the observed iodide efflux may be in part due to spontaneous channel

activity, HEK293 cells expressing α3β2γ2 GABAA receptors were investigated using the

patch-clamp method. Administration of 100 µM picrotoxin to the cells showed small out-

ward currents in cells expressing the receptors with an agonist-independent receptor activity

of 9.04 ± 1.75 % of total response capacity (maximal GABA response + response to picro-

toxin = 100 %, n = 8; Figure 4.14). 1

2 sec

500 pA

GABA PTX

A B

Figure 4.14: Representative current traces evoked by 100 µM GABA (A) and 100 µM picro-toxin (B) in HEK293 cells expressing α3β2γ2 GABAA receptors.

4.2.3 Optimisation of a YFP-based screening assay

A third possible assay option, a YFP mutation with increased chloride and iodide sensi-

tivity, was tested using a fluorescence plate reader. HEK293 cells transiently transfected

with the mutated YFP-H148Q/I152L as well as α3β2γ2 GABAA receptors were seeded into

96-well plates. On the following day, cells were incubated in NaCl bath solution an hour

before the assay. Afterwards, the microtitre plates were put into a fluorescence plate reader

and NaI test solution only or together with 100 µM GABA were injected into the wells and

changes in the fluorescence signal were measured. The latter resulted in a clear fluorescence

decrease, which could be prevented by injecting 100 µM GABA together with 1 mM picro-

toxin (Figure 4.15 A). These results indicated that α3β2γ2 GABAA receptor activity, elicited

by a high agonist concentration could be detected with this assay. It was, therefore, further

1During this set of experiments GABA concentration-response relations were also examined for the α3β2γ2receptors (see Figure B.6, Appendix B).

103


tested in HEK293 cells transfected with YFP-H148Q/I152L and α3β2 or α3β2ε GABAA

receptors. Cells expressing the former receptors also exhibited a fluorescence quench, which

was not seen when picrotoxin was applied. However, the fluorescence quench was only

12.41± 3.55 % (n = 9) compared to 54.62± 4.14 % (n = 15) obtained from cells expressing

α3β2γ2 receptors (Figure 4.15 B). When examining cells expressing α3β2ε receptors, the

fluorescence signal did not show distinct changes after adding the agonist or channel blo-

cker (n = 9; Figure 4.15 C). From the results obtained, it was suggested that the assay could

probably be used for certain GABAA receptor combinations, such as the α3β2γ2 subtype.

However, the assay seemed unsuitable for the receptor subtype containing the ε subunit,

since no receptor activity could be detected.

Since the various subunit combinations resulted in different assay outcomes, i.e. different

percentage of fluorescence quench, it was tested whether another α subunit, namely the

human α1 subunit, would improve the fluorescence readings when combined with β2 and

ε subunits. HEK293 cells expressing α1β2 and α1β2ε, respectively, were examined as

described above and produced some interesting results. Cells expressing the α1β2 receptor

showed a 39.4 ± 3.32 % quench (n = 9) which was significantly different compared to

the results obtained with the α3β2 receptor (p > 0.0001, unpaired t-test; Figure 4.16 A).

Similarly, a fluorescence quench of 24.65± 2.13 % (n = 9) was measured in cells expressing

α1β2ε receptors after adding 100 µM GABA to the wells (Figure 4.16 B). This could be

prevented by adding a high concentration of picrotoxin.

In order to assess the quality of the assay, the Z’-factor for each receptor subtype was cal-

culated as described in Section 2.6.3.1 (Equation 2.7, Chapter 2). Although the fluorescence

quench for α3β2γ2 receptors was the highest obtained, the Z’-factor was 0.48. Further Z’-

factors were -0.16 for α3β2, 0.58 for α1β2 and -0,01 for α1β2ε.

Next, GABA concentration-response relations were examined in cells expressing α3β2γ2,

α1β2 and α1β2ε receptors, respectively. Figure 4.17 displays that different subunit combina-

tions resulted in different GABA concentration-response curves, when investigating GABA

concentrations ranging from 30 nM to 3 mM. α3β2γ2 receptors were most sensitive to

104


A

40

60

80

100

120

5 10 15 20 25 30 35BF

NaI (n = 9)NaI + 100 µM GABA (n = 9)NaI + 100 µM GABA + 1mM PTX (n = 9)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

B

C

40

60

80

100

120

5 10 15 20 25 30 35BF


Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35BF


Time [sec]

% N

et fl

uo

resc

ence

sig

nal

Figure 4.15: Net fluorescence signal of mutant YFP-H148Q/I152L in HEK293 cells expres-sing α3-subunit-containing GABAA receptors.Fluorescence was measured for 30 seconds after addition of NaI test solution only or together with100 µM GABA or with 100 µM GABA and 1 mM picrotoxin in cells expressing α3β2γ2 (A), α3β2(B), α3β2ε (C) GABAA receptors. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence was normalised by setting the averaged baseline fluorescence to 100 %and fluorescence changes (after adding solutions) were calculated as percentage of baseline fluores-cence.

105


40

60

80

100

120

5 10 15 20 25 30 35

NaINaI + 100 µM GABANaI + 100 µM GABA + 1mM PTX

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A

40

60

80

100

120

5 10 15 20 25 30 35BF

NaINaI + 100 µM GABANaI + 100 µM GABA + 1mM PTX

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

B

Figure 4.16: Net fluorescence signal of mutant YFP-H148Q/I152L in HEK293 cells expres-sing α1-subunit-containing GABAA receptors.Fluorescence was measured for 30 seconds after addition of NaI test solution only (n = 9) or togetherwith 100 µM GABA (n = 9) or with 100 µM GABA and 1 mM picrotoxin (n = 9) in cells expressingα1β2 (A) and α1β2ε (B) GABAA receptors. Data correspond to means ± SEM; n states the numberof wells measured. Net fluorescence was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

106


GABA (EC50 3.37 ± 0.42 µM, n = 14). α1β2 receptors had an EC50 of 8.47 ± 0.48 µM

(n = 10) and α1β2ε receptors where least sensitive with an EC50 of 17.1 ± 5.4 µM (n = 9).

The Hill slope for α1β2ε receptors was lower (1.17 ± 2.70) compared to those obtained for

α3β2γ2 (2.26 ± 0.36) and α1β2 (2.57 ± 0.14) receptors. However, the difference was not

significant.

0

25

50

75

100

125

10 20 30

NaI100 nM GABA300 nM GABA1 µM GABA3 µM GABA10 µM GABA30 µM GABA100 µM GABA300 µM GABA1 mM GABA3 mM GABA

BF

Time [sec]

% N

et fl

uore

scen

ce q

uenc

h

-9 -8 -7 -6 -5 -4 -3 -20

10

20

30

40

50

60

α3β2γ2

α1β2α1β2ε

log [GABA, M]

% N

et

flu

ore

sce

nce

qu

en

ch

0

25

50

75

100

125

10 20 30

NaI

100 nM GABA300 nM GABA1 µM GABA3 µM GABA10 µM GABA30 µM GABA100 µM GABA300 µM GABA1 mM GABA

BF

30 nM GABA

Time [sec]

% N

et fl

uore

scen

ce q

uenc

h

60

80

100

120

10 20 30

NaI

100 nM GABA300 nM GABA1 µM GABA3 µM GABA10 µM GABA30 µM GABA100 µM GABA300 µM GABA1 mM GABA

BF

30 nM GABA

Time [sec]

% N

et fl

uore

scen

ce q

uenc

h

A B

C D

Figure 4.17: GABA concentration-response relations examined for GABAA receptors ex-pressed in HEK293 cells using the YFP-based screening assay.Time-course showing GABA-induced net fluorescence quench using increasing concentrations ofGABA in cells expressing (A) α3β2γ2, (C) α1β2 and (D) α1β2ε receptor subtypes. (B) GABAconcentration-response curves of α1β2 (n = 10), α1β2ε (n = 9) and α3β2γ2 receptor subtypes(n = 14). Data correspond to means ± SEM; n states the number of wells measured. EC50 va-lues and Hill coefficients are stated in the text. Curves were fitted using a non-linear regression fitwith variable slopes as described in Section 2.6.3.

1 µM of the benzodiazepine flunitrazepam (together with the EC30 concentration of GABA)

was applied to cells expressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors. Results

were compared to the action of an EC30 concentration of GABA alone (Figure 4.18). In

cells exposed to 2.45 µM GABA alone, the fluorescence drop was 30.3 ± 5.88 % compa-

107


red to the baseline fluorescence, whereas it was 39 ± 5.76 % in cells exposed to 2.45 µM

GABA with 1 µM flunitrazepam, demonstrating the positive modulation of the receptor by

the benzodiazepine.

40

60

80

100

5 10 15 20 25 30 35

NaI100 µM GABA2.45 µM GABA2.45 µM GABA + 1 µM FLZ

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 4.18: Fluorescence signal change after addition of 1 µM flunitrazepam to HEK293cells expressing α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after addition of NaI test solution only (n = 8) or togetherwith 100 µM GABA (n = 12), 2.45 µM GABA (n = 12) or 2.45 µM GABA and 1 µM flunitrazepam(n = 9) in cells expressing α3β2γ2 receptors. Concentrations are final concentrations per well. Datacorrespond to means ± SEM; n states the number of wells measured. Net fluorescence was norma-lised by setting the averaged baseline fluorescence to 100 % and fluorescence changes (after addingsolutions) were calculated as percentage of the baseline fluorescence.

4.3 Discussion

4.3.1 Clomeleon-based screening assay

The results presented in this chapter highlight that chloride-ion channel screening is challen-

ging. Several methods had to be tested to find a suitable assay technique. The different assays

were first tested and optimised with cells expressing α3β2 or α3β2γ2 receptors, since those

subtypes did not exhibit constitutive receptor activity. The latter meant a further challenge

for the assay setup when using ε-containing receptors.

The initial plan was to set up a cell-based system to screen for compounds that modulate the

GABAA receptor using the chloride-ion indicator protein Clomeleon (Kuner & Augustine,

2000). It had been shown that Clomeleon is suitable for chloride concentration measure-

108


ments in neurons of the hippocampus as well as transgenic mice (Kuner & Augustine, 2000;

Duebel et al., 2006). However, Clomeleon had not been used in a cell-based screening as-

say to scan GABAA receptors for modulating compounds as attempted in this project. The

excitation and emission spectra obtained from HEK293 cells expressing Clomeleon were

in consent with the data published by Kuner & Augustine (2000). Therefore, Clomeleon

was next expressed with α3β2 GABAA receptors and the cells were seeded into 96-well

microtitre plates. Using a fluorescence plate reader, the fluorescence signal was monitored

before and after adding a high concentration of the agonist GABA to the cells. However,

data obtained did not result in the expected fluorescence shift, with no or occasionally small,

non-significant signal changes compared to baseline fluorescence. This slight fluorescence

reductions could be measured in few wells at both emission wavelengths of Clomeleon. It

was suggested that these changes were not Clomeleon-specific, but rather an artefact of the

assay setup. This is because fluorescence was only read at a small-sized area of the wells. In

some cases, cells probably got detached from the plate when injecting the agonist-solution

which lead to reduction of cells in the area of the signal detection, decreasing the fluores-

cence signal. It was supposed that chloride-sensitivity of Clomeleon was too low to detect

receptor mediated changes in the cells even at high agonist concentrations and, thus, no

chloride-specific signal changes could be detected. In fact, the physiological range of chlo-

ride concentrations in different cell types lies between 3 to 60 mM (Rohrbough & Spitzer,

1996; Kuner & Augustine, 2000; Berglund et al., 2008; Markova et al., 2008; Bregestovski

et al., 2009), but sensitivity of Clomeleon to chloride is much lower (KApp > 160 mM).

Hence, it was concluded that Clomeleon is not suitable for the intended assay type due to its

low chloride-sensitivity.

Since the initially planned screening assay setup did not produce the desired outcome and

proved unsuitable for screening compounds, an alternative screening technique had to be

found. There were two possible methods, which seemed suitable due to their relatively easy

setup and no need for specialised measuring devices: a colorimetric-flux assay and a YFP-

based screening assay.

109


4.3.2 Colorimetric iodide-flux assay

The next assay tested was the colorimetric iodide-flux assay described by Tang & Wildey

(2004, 2006) which was based on a modified SK method. For this, HEK293 cells expres-

sing the α3β2γ2 receptor were loaded with iodide. Subsequent application of GABA to

the cells caused the ion channel receptors to open, triggering iodide efflux through the ion

channel receptors into the supernatant solution. By determining the intra- and extracellular

iodide concentrations, it would be possible to evaluate GABAA receptor activity. No or low

concentrations of GABA result in low receptor activity, hence, the amount of iodide in the

cells should be high, resulting in low OD405 values, whereas applying high concentrations

of GABA results in high receptor activity. This causes most of the iodide to flow out of the

cells due to the electrochemical gradient, causing a low intracellular iodide concentration

and, hence, high OD405 values (see Equation 4.1). When determining extracellular iodide

concentrations, OD405 values should be low at high receptor activity and high at low or no

receptor function.

Standard curves obtained after increasing incubation times displayed a clear reduction of the

OD405 readings over time, confirming the relatively short stability of the mixture of iodide-

containing samples and detection buffers reported by Tang & Wildey (2004). The minimal

detection limit estimated from the standard curve was higher than the one described by Tang

& Wildey (2004). This may arise from the fact that the OD values obtained here were lower

and probably due to well-to-well variations, since the 3 × SD is comprised to evaluate the

detection limit. The assays linear range was determined from the NaI standard curves exhi-

biting a range of about 300 nM to 3 µM after 15 minutes assay incubation time or after 30

minutes between 100 nM and 1 µM. The lower value determined after 30 minutes was simi-

lar to the value presented by Tang & Wildey (2004), who reported a linear range of 130 nM

to 13 µM after 30 minutes incubation time. However, the upper linear range limit gained

was 10-fold lower than the one described by Tang & Wildey (2004), who showed a 100-fold

molar linear range of the method. Thus, the SK assay detection range obtained here was

relatively small compared to the data represented by Tang & Wildey (2004). The reasons for

these variations are not clear, since detection buffers used were prepared as described (Tang

110


& Wildey, 2004, 2006). There may have been slight variations in NaI solutions or detec-

tion buffers due to fluctuation in measuring, weighing or pipetting, however, it is unlikely

that these deviations are responsible for the indicated differences, especially since standard

curves were repeated a few times using different solutions and buffers.

Next, the assay was tested in HEK293 cells transiently transfected with α3β2γ2 GABAA re-

ceptors using 96-well microtitre plates. As mentioned above, high OD405 values indicate a

low iodide concentration inside the cells, whereas low OD405 readings represent high intra-

cellular iodide concentrations. However, readings obtained from cell lysates after applying

ascending GABA concentrations to the cells were rather random and did not show specific

ion flux or GABA concentration-related patterns. Furthermore, untransfected control cells

exhibited similar OD405 readings as transfected cells, although they did not express the exo-

genous GABAA receptors. Cell supernatants were also subject to the modified SK reaction.

The supernatant solution was iodide free before adding it to the cells and should contain

different concentrations of iodide depending on the level of receptor activation and, hence,

the iodide efflux level. OD405 readings obtained from undiluted supernatants were low,

indicating high iodide concentrations in the wells and confirming that iodide effused from

the cells. Due to the small detection window, supernatant solutions were diluted in order to

detect differences in the iodide concentrations per well that could distinguish GABAA re-

ceptor activity in the differently treated wells. However, the data obtained were not GABA

concentration-related but random in transfected as well as untransfected cells. The fact that

readings from untransfected control cells were similar to the ones obtained from cells tran-

siently expressing GABAA receptors indicated a very low signal-to-noise ratio. This could

be due to an imperfect removal of extracellular iodide after loading the cells or due to a

continuous leakage of iodide from the cells before receptor stimulation (Zheng et al., 2004).

However, cells were washed three to four times after the loading step, which most likely

excludes the former possibility. A non-specific iodide leakage through the cell membrane

could happen via passive diffusion, chloride pumps or spontaneous active GABAA receptors

and may be one possible explanation for the low signal-to-noise ratio (Smith et al., 2001).

After repeating the assay a few times with no clear concentration-dependent effects, dif-

111


ferent assay components were varied and tested in order to improve and optimise the assay

and to increase the signal-to-noise ratio. The agonist incubation time is an important factor

for ideal flux responses that are dependent on the balance between ion channel activation

and desensitisation (Smith et al., 2001). Also, iodide-leakage that may play a part in the

low signal-to-noise ratio detected, could probably be reduced with shorter agonist incuba-

tion time which should then result in more accurate readings. Different GABA incubation

times between ten seconds to five minutes were examined to optimise iodide efflux from the

cells. Data obtained were compared with control data from DPBS incubation at the different

time points in order to subtract non-specific iodide efflux from efflux induced by GABA.

Supernatant data from certain time points differed significantly between GABA- and DPBS-

treated cells. However, the test results were inconsistent and not reproducible. Also, often

GABA-treated and control cell OD405 values were similar or occasionally flux from control

cells seemed to be significantly more than from GABAA receptor expressing cells, although

the opposite was expected. Due to the discrepancies obtained throughout the data, no optimal

incubation time could be determined and it was decided to use the incubation time suggested

by Tang & Wildey (2004) while other test variables were tried to be optimised instead. The

question remained, why could the signal-to-noise ratio not be improved by different agonist

incubation times and why were the results so difficult to reproduce?

It was also tested whether changes in cell density would improve the assay outcome. In pre-

vious experiments, iodide concentrations in the supernatants were relatively high whereas

it was low in cell lysates, which could indicate high receptor activity, which may have not

been suitable for the small detection range of the assay. Thus, it was examined whether lower

cell numbers would improve intra- and extracellular iodide concentrations for better OD405

readouts. At all cell densities tested OD405 values were very low for supernatant solutions

and no differences could be seen between the different cell densities without further super-

natant dilutions. However, results obtained from cell lysates with 25 × 103 or 17.5 × 103

cells per well showed significant differences in OD405 readouts between different GABA

concentrations, but no concentration-related pattern could be detected. Cell lysates of wells

containing 12.5× 103 cells per well, however, exhibited a significant lower reading for 1 µM

112


GABA compared to the three higher concentrations tested. Although this could have been

a concentration-related difference, it was assumed, considering the high error bars obtained,

that this was part of the variation seen throughout the plates and, hence, overlooked and not

further examined. By hindsight, it would have been perspicacious to further decrease the

number of cells. However, since the results were again not well-defined and Tang & Wildey

(2004) used 2.5 × 104 WSS-1 cells per well, it was decided to continue with cell densi-

ties used by Tang & Wildey (2004), because WSS-1 cells, which stably express GABAA

receptors, were derived from HEK293 cells (Wong et al., 1992).

Due to the fact that all assay optimisation attempts failed to improve the assay by producing a

well-defined signal window, the functionality of the expressed GABAA receptors was investi-

gated by applying the competitive antagonist bicuculline and the channel blocker picrotoxin.

Again, OD405 data obtained for cell lysates were relatively high; only few differences were

seen between the distinctly treated transfected and untransfected cells. However, results ob-

tained from the supernatants were interesting. In all but the picrotoxin-treated wells, OD405

values were very low indicating a high iodide efflux from the cells. However, picrotoxin-

treated cells, transfected as well as untransfected, exhibited significantly higher OD405 va-

lues, indicating a low concentration of iodide and, thus, blockage of the iodide-efflux from

the cells. Since picrotoxin is a chloride channel blocker, the results indicated that mainly

chloride channels were responsible for the iodide leakage detected in previous experiments,

rather than chloride pumps or spontaneously active channels. Although, patch-clamp expe-

riments using α3β2γ2 receptor expressing HEK293 cells revealed small constitutive activity

of the expressed GABAA receptors. It is not clear why iodide concentrations were very high

in cell supernatants containing bicuculline, since bicuculline is a receptor antagonist.

The colorimetric iodide-flux assay setup exhibited difficulties that could not be successfully

amended when trying to optimise certain assay parameters. An important feature of com-

pound screening assays is the assays reliability as well as a good information content and a

robust assay setup (Xu et al., 2001), which could not be achieved with this assay format using

HEK293 cells. There were three major issues, the variability between replicates, the random

readings between different treatments and the fact that transfected as well as untransfected

113


cells had similar results. It is likely that several factors contributed to the assays failure. One

is that the assay’s detection range was very small. Therefore, already small iodide carry-

overs could completely invalidate an experiment. Great care was taken to avoid carry-overs,

but since the many pipetting steps involved were carried out manually, this may have been

one contributing factor for the assays random readings and poor reproducibility between ex-

periments. Another cause could have been the variable receptor population among the cells

and especially between the different experiments due to variations in transient transfection

efficacy. It is proposed that transient transfection efficiencies vary between 10 to 50 %, de-

pending on cell type, cDNA and transfection reagent (Witchel et al., 2002; Thomas & Smart,

2005; Gilbert et al., 2009b) and it is likely that many cells in the wells were untrasfected.

This heterogeneous cell population also likely added to the low signal-to-noise ratio seen

throughout (Lynch, 2005). In comparison, Tang & Wildey (2004) demonstrated the assays

functionality and reproducibility in cells expressing GABAA receptors stably and even then

the Z’-factor that determines the assay’s quality lay below 0.5, indicating a relatively small

signal window that is not ideal for the assay’s application in HTS (Zhang et al., 1999; Tang

& Wildey, 2004). Even when considering a relatively low transfection efficacy, one question

remained: why was the iodide efflux in cells expressing the GABAA receptors and control

cells similar? It was suggested that agonist-independent iodide fluctuations, due to passive

diffusion, active chloride pumps and spontaneous receptor activity (at least in transfected

cells) partly contributed. A study by Galietta et al. (2001b) showed that the assay tempe-

rature can significantly influence basal halide permeability. In most cell types, non-specific

halide permeability is significantly reduced when performing an assay at 37°C compared to

room temperature, where it was relatively high. This likely explains the high iodide-efflux

measured from untransfected control cells. Nevertheless, results obtained from applying pi-

crotoxin to the cells indicated that the observed iodide-efflux blockade probably occurred

via chloride channels. HEK293 cells endogenously express β3, γ3 and ε GABAA receptors

(Thomas & Smart, 2005). They also express GlyR β subunits. However, the latter do not

form functional homomeric channels (Lynch, 2009). There have been reports that endoge-

nously expressed GABAA receptors can trigger GABA-activated currents (Ueno et al., 1996;

114


Thomas & Smart, 2005), however, patch-clamp studies on untransfected cells performed in

this project did not detect GABA-activated currents. It is also worth mentioning that pre-

incubation of cells with iodide-containing buffers can affect cell viability (Galietta et al.,

2001b), which may have also contributed to the assay’s outcome. Although the reasons for

the assay’s failure are not entirely clear, it was clearly not reliable or robust in the current as-

say format using transiently transfected HEK293 cells. Therefore, a further alternative plate

reader assay was examined.

4.3.3 YFP-based screening assay

The final screening assay setup tested was based on the iodide-sensitivity of the mutant

YFP-H148Q/I152L that can be used to determine iodide influx into cells via membrane

proteins, such as the GABAA receptor (Galietta et al., 2001a; Kruger et al., 2005; Gilbert

et al., 2009b). By using this assay format with HEK293 cells expressing YFP-H148Q/I152L

and α3β2γ2 receptors, it could be demonstrated that it was possible to measure changes

in the fluorescence signal due to iodide influx into the cells upon GABAA receptor activa-

tion. The fluorescence change obtained was around 50 % and, thus, similar in magnitude

as seen in other studies using GABAA receptors (Kruger et al., 2005; Gilbert et al., 2009b).

Gilbert et al. (2009b) detected a larger fluorescence quench with α1β1 receptors compared

to changes obtained when adding a γ subunit to the receptor complex, whereas here the

largest fluorescence quench was obtained from γ2-containing receptors compared to α1β2

receptors. Several studies have investigated kinetic properties of different receptor isoforms,

concluding that GABAA receptor channel gating is dependent on receptor subunit composi-

tion (Verdoorn, 1994; Gingrich et al., 1995; Lavoie et al., 1997; Haas & Macdonald, 1999).

It was shown that the γ2L subunit confers longer duration openings and longer bursts of

opening to a receptor subtype, compared to αβ-containing receptors (Haas & Macdonald,

1999), which could explain the higher fluorescence quench obtained. It was proposed that

the fluorescence variation seen for the different receptor subtypes could be due to the distinct

kinetics and gating properties of the receptor subtypes investigated, that conducted different

levels of iodide influx. A further factor may be differences in transfection efficiency as shown

115


by Gilbert et al. (2009b). It is unclear why α3β2ε receptors could not trigger sufficient iodide

influx to reduce the level of fluorescence inside the cells whereas a reduction in fluorescence

could be measured in cells expressing α1β2ε receptors. The spontaneous receptor activity is

most likely one factor contributing to the lower fluorescence quench achieved in ε-containing

receptors. Furthermore, Whiting et al. (1997) reported a rapid desensitisation rate in α1β1ε

receptors expressed in HEK293 cells. This would explain a lower and less sufficient iodide

influx compared to other receptor subtypes exhibiting slower deactivation and desensitisation

rates.

Before testing different GABA concentrations and compounds using the assay setup, its qua-

lity was evaluated and results obtained for different GABAA receptor subtypes were compa-

red among each other by determining the assays’ Z’-factors. The Z’-factor incorporates the

assay’s signal dynamic range as well as data variation between replicates, without interfe-

rence of test compounds (Zhang et al., 1999). Results obtained indicated that the setup of

the assay is acceptable when using cells expressing both α1β2 (Z’ = 0.58) and α3β2γ2 re-

ceptors (Z’ = 0.48; Inglese et al., 2007a). However, in cells expressing α1β2ε receptors the

signal window was relatively small that positive and negative signal variation bands overlap-

ped slightly, which caused a negative Z’-factor. Although the assays outcome was improved

and a fluorescence quench could be detected when substituting the α3 subunit with the α1

subunit, the assay’s quality using cells expressing the ε subunit was not ideal. Hence, it

was decided to screen the compound library on both α3β2γ2 receptors as well as on α1β2ε

receptors. Cells expressing the latter were used despite the small signal window, as it was

proposed that positively modulating as well as inhibiting compounds could still be identified

by using this method.

Next, GABA concentration-response relations were investigated. By this means it was pos-

sible to compare the results of the assay with data obtained from electrophysiological me-

thods as well as to determine a suitable submaximal GABA concentration for the screening

assay. The findings gained were very different to the ones obtained from electrophysiolo-

gical recordings (see Section 3.2.1, Chapter 3). EC50 values from cells expressing α3β2γ2

receptors were lower than those from cells expressing α1β2 and α1β2ε receptors; the latter

116


exhibited the highest values. Using electrophysiological recordings in oocytes it was shown

that the ε subunit confers high agonist sensitivity to the receptor (Section 3.2.1, Chapter 3),

which was also shown in several other studies (Whiting et al., 1997; Neelands et al., 1999;

Maksay et al., 2003; Ranna et al., 2006). Furthermore, in case of α3β2γ2 receptors, the

EC50 value of the plate reader assay was lower compared to the ones obtained from oocytes

(39.37 µM) or HEK293 cells (14.65 µM; Figure B.6, Appendix B) using electrophysiological

methods, whereas the EC50 value for α1β2ε receptors was higher than the one obtained in

oocytes (3.55 µM; Figure B.5, Appendix B). Kruger et al. (2005) measured concentration-

response relations in cells expressing the mutant YFP and α1 GlyR using an imaging-based

assay system as well as a fluorescence plate reader. It was reported that the EC50 values ob-

tained with the latter method were five-times higher compared to the former technique. Due

to the high deviation seen with the fluorescence plate reader assay, compared to the other

systems tested, the reliability of the assay format was doubted (Kruger et al., 2005). Howe-

ver, it is not surprising that the results obtained here vary, considering the variation between

the different measuring techniques. The YFP-based assay was performed in a plate reader

that measured values from the whole cell population in the wells, the measuring time was

extended and an indirect, secondary effect (the YFP quench) was detected. In comparison,

electrophysiological recordings are performed on single cells for a short time and current

flow across the membrane upon receptor activation can be measured directly (Wafford et al.,

2009). Furthermore, mutant YFP variants exhibited relatively slow chloride association and

disassociation kinetics (iodide binding kinetics were not shown; Galietta et al., 2001a) and

Colquhoun (1998) stated that to gain precise concentration-response curves fast agonist ap-

plication is vital due to the fast desensitisation rates of ion channels.

By comparing the fluorescence signals of rising concentrations of GABA it was decided to

use EC30 GABA concentrations when screening the compounds. This is because fluores-

cence signals obtained for EC30 GABA concentrations could be clearly distinguished bet-

ween positive (high GABA concentration) and negative controls (NaI test solution only or

GABA (EC30) with picrotoxin). This was in consent with work done by Wafford et al. (2009),

who also used EC30 concentrations, although in a different screening assay type (FLIPR).

117


Benzodiazepines are known to positively modulate GABAA receptors and flunitrazepam po-

tentiated submaximal GABA concentrations in α3β2γ2 receptors (Section 3.2.4, Chapter 3).

Using the YFP-based assay setup it could be demonstrated that the potentiating effect by

flunitrazepam could also be detected with the fluorescence plate reader.

The different tests performed using the YFP-based screening assay demonstrated that the

assay was suitable to detect GABA-induced receptor activity when using cells expressing

α3β2γ2 and, to a certain degree α1β2ε receptors. It was also shown that the fluorescence

signal change was GABA concentration-dependent and that the benzodiazepine-induced

potentiation of a submaximal GABA concentration could be detected in cells expressing

α3β2γ2 receptors. This confirmed the suitability of the assay for pharmacological investi-

gations.

4.3.4 Summary

The results presented in this chapter clearly display the difficulties associated with setting

up a chloride-ion channel screening assay. The data show that a specific assay setup, that

works for a certain cell line and chloride channel or GABAA receptor subtype, not necessa-

rily works equally well with other receptor (sub)types or cell lines, even after optimisation

steps. For example, attempts were made to optimise many parameters when setting up the

iodide-flux assay. However, at the end it could be concluded that the assay format described

by Tang & Wildey (2004) was unsuitable when using transiently transfected HEK293 cells

expressing GABAA receptors. Nevertheless, an acceptable assay to screen certain GABAA

receptor subtypes was found. However, there were certain limitations and difficulties when

investigating spontaneously open ε-containing receptors and special care should be taken

when evaluating screening results, due to the small signal detection window of the assay.

118

Chapter 5

Screening a chemical compound library

using the mutant YFP-H148Q/I152L

5.1 Introduction

Not much is known about the ε-subunit-containing GABAA receptor. Studies conducted pre-

viously could not determine a subunit-specific pharmacological profile and its functions in

vivo are still not known. It was proposed that ε-containing receptors have a specific modula-

tory site, similar to the benzodiazepine binding site. For these reasons, the next aim of this

study was to screen an ε-containing GABAA receptor subtype against a chemical compound

library and, with luck, to find compounds that modulate receptor activity in the presence of

GABA. ε-subunit-selective compounds will be of great use to elucidate the subunit’s role

in vivo by indicating ε-specific pharmacological effects or possibly a specific phenotype.

Furthermore, amino acids important for the compound’s binding site could be identified by

using molecular approaches, which would further help to study the subunit’s role. Moreover,

a novel lead compound could be a possible future drug acting on the receptor, for example

to treat AD (Section 1.4, Chapter 1).

Several cell-based screening assays are available, such as flux or YFP-based assays, and

many aspects have to be considered when developing and setting up a suitable approach, for

example assay format and detection instrumentation (Inglese et al., 2007b). In the previous

119

Chapter 5. Screening a chemical compound library

chapter (Chapter 4), a number of screening assay formats were tested for their suitability to

screen compounds against certain GABAA receptor subtypes. A cell-based screening assay

using the mutant YFP-H148Q/I152L gave robust data when using HEK293 cells expressing

α3β2γ2 receptors. Results obtained for ε-containing receptors exhibited a relatively small

signal dynamic range, limiting the assay’s suitability. However, it was proposed that effects

of inhibiting and enhancing compounds could be distinguished from submaximal GABA-

control responses. Due to the lack of other alternative assay formats, the YFP-based assay

was applied to screen compounds and search for α3β2γ2 and α1β2ε GABAA receptor sub-

type modulators. By this means, the effects and specificity of hit compounds of the two

receptor subtypes could be compared, and ε-subunit-specific compounds could be identified

more easily. A similar study by Wafford et al. (2009) sought compounds specific to the δ

subunit and also compared effects of the hit compounds with their effect on αβγ receptors.

The compound library purchased (Maybridge HitFinder™Collection) comprised of drug-

like molecules with a broad range of structural motifs and biological space. The collection is

rich in diverse pharmacophores, and compounds follow Lipinski guidelines (Lipinski et al.,

2001). The Maybridge HitFinder™Collection consists of 14,400 compounds in total, of

which a small selection of 800 compounds has been tested here. Due to the fact that nothing

is known about specific compound-binding sites on the ε subunit, it was important to screen

compounds with great structural diversity, which is a key factor concerning chemical libra-

ries, especially since the drug target site is unknown (Gillet, 2008). It is likely that similar

compound structures would exhibit similar activities. Therefore, chances to find chemical

leads are enhanced when using a broad range of structural motifs. Finally, serendipity is

also an important factor to find an ε-subunit-selective compound among the 800 compounds

investigated in this study.

The aim of this chapter was to identify subtype-selective compounds modulating α1β2ε

and α3β2γ2 GABAA receptor subtypes by performing preliminary screens using the mu-

tant YFP-H148Q/I152L. Afterwards, it was aimed to confirm the identity of hit compounds

and further examine the compounds in more detail using the two-electrode voltage-clamp

technique.

120


5.2 Results

5.2.1 Preliminary screening

Test compounds were screened against both α3β2γ2 and α1β2ε GABAA receptor subtypes.

Note that compounds were designated ’TS-’ and in addition numbered (1-41). The preli-

minary screen was carried out using relatively high concentrations of the compounds (final

concentrations between 50.3 and 59.29 µM). 1 µl of a compound (in DMSO, see Sections

2.7.1 and 2.7.2, Chapter 2) was added to each well containing 50 µl NaCl bath solution,

and an EC30 concentration of GABA in NaI test solution was subsequently injected into

the wells. By this means, compounds were identified that exhibit positive or negative mo-

dulatory effects on the receptor subtypes by exhibiting a reduced or increased fluorescence

quench compared to cells treated with EC30 GABA concentrations only or cells subjected

to GABA (EC30) and 0.5 % DMSO (compound vehicle only negative control). The results

acquired indicated that the majority of test compounds had no effect on the receptors. In the

case of α1β2ε receptors 73.74 % of all compounds were inactive; Figure 5.1 represents a

selection of negative screening results. In comparison, 87.31 % of test compounds had no

effect on α3β2γ2 receptors (Figure 5.2).

Figures 5.3 and 5.4 display representative results of test compounds that showed an inhibitory

(5.3 A, B and 5.4 A) or positive (5.3 C and 5.4 B, C) effect on receptor activation. In the case

of α1β2ε receptors, there were 209 compounds at this concentration range of which 151

inhibited and 58 increased the YFP fluorescence quench. In comparison, 101 compounds

seemed to evoke an explicit effect on cells expressing α3β2γ2 receptors of which 10 were

inhibiting and 91 were increasing the YFP quench.

121


70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA

8.26 µM GABA + 1 mM PTX8.26 µM GABA

8.26 µM GABA + 0.5 % DMSO8.26 µM GABA + compound TS-25

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

B

C

70

80

90

100

110

120

5 10 15 20 25 30 35BF

100 µM GABA



Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

D

E

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

F

Figure 5.1: Sample negative results from the preliminary screening of HEK293 cells expres-sing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA (n = 3), 8.26 µM GABA(EC30) and 1 mM picrotoxin (n = 3), 8.26 µM GABA and 0.5 % DMSO (n = 3), 8.26 µM GABAalone (n = 3) or 8.26 µM GABA with test compound (n = 1): (A) compound TS-25 (54.37 µM), (B)compound TS-26 (53.81 µM), (C) compound TS-27 (53.92 µM), (D) compound TS-28 (56.72 µM),(E) compound TS-29 (52.52 µM), (F) compound TS-30 (58.47 µM). Compound concentrations arefinal concentrations per well after injection of GABA. Data correspond to means ± SEM; n statesthe number of wells tested. Net fluorescence was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

122


D

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA


2.45 µM GABA + 0.5 % DMSO

2.45 µM GABA + compound TS-34

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA




BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

E F

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA




BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35BF

100 µM GABA



Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C

Figure 5.2: Sample negative results from the preliminary screening of HEK293 cells expres-sing mutant YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA (n = 3), 2.45 µM GABA(EC30) and 1 mM picrotoxin (n = 3), 2.45 µM GABA and 0.5 % DMSO (n = 3), 2.45 µM GABAalone (n = 3) or 2.45 µM GABA with test compound (n = 1): (A) compound TS-31 (54.58 µM), (B)compound TS-32 (54.31 µM), (C) compound TS-33 (54.88 µM), (D) compound TS-34 (54.08 µM),(E) compound TS-35 (55.66 µM), (F) compound TS-36 (57.12 µM). Compound concentrations arefinal concentrations per well after injection of GABA. Data correspond to means ± SEM; n states thenumber of wells measured. Net fluorescence was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

123


70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 2)

8.26 µM GABA + 1 mM PTX (n = 3)8.26 µM GABA (n = 2)

8.26 µM GABA + 0.5 % DMSO (n = 3)8.26 µM GABA +compound TS-31 (n = 1)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 2)


8.26 µM GABA + 0.5 % DMSO (n = 3)8.26 µM GABA + compound TS-38 (n = 1)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 3)


8.26 µM GABA + 0.5 % DMSO (n = 3)

8.26 µM GABA + compound TS-37 (n = 1)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A

B

C

Figure 5.3: Representative positive screening results for HEK293 cells expressing mutantYFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA, 8.26 µM GABA and1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO, 8.26 µM GABA alone or 8.26 µM GABA withtest compound: (A) compound TS-31 (54.58 µM), (B) compound TS-37 (55.04 µM), (C) compoundTS-38 (55.24 µM). Concentrations are final concentrations per well after injection of solutions. Datacorrespond to means± SEM; n states the number of wells tested. Net fluorescence was normalised bysetting the averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions)were calculated as percentage of the baseline fluorescence.

124


40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA (n = 2)


2.45 µM GABA + 0.5 % DMSO (n = 3)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

60

80

100

120

5 10 15 20 25 30 35

100 µM GABA (n = 2)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A

B

C

50

60

70

80

90

100

110

120

130

5 10 15 20 25 30 35

100 µM GABA (n = 2)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.4: Representative positive screening results for HEK293 cells expressing mutantYFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA, 2.45 µM GABA and1 mM picrotoxin, 2.45 µM GABA and 0.5 % DMSO, 2.45 µM GABA alone or 2.45 µM GABA withtest compound: (A) compound TS-39 (55.32 µM), (B) compound TS-40 (56.32 µM), (C) compoundTS-41 (55.82 µM). Concentrations are final concentrations per well after injection of solutions. Datacorrespond to means± SEM; n states the number of wells tested. Net fluorescence was normalised bysetting the averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions)were calculated as percentage of the baseline fluorescence.

125


The number of compounds which had an effect on the two GABAA receptor subtypes was

very high during the initial screen. Therefore further screens were performed by gradually

decreasing the concentration of the test compounds in order to identify the most potent com-

pounds. In a second screen all compounds which indicated positive receptor modulation or

inhibition at the initial high concentrations were tested again with a tenfold diluted concen-

tration (between 5.03 and 5.93 µM). Following this scan, the number of active compounds

on α1β2ε receptors decreased to 95 (86 inhibiting and 9 enhancing the receptor activity). 32

compounds still showed an effect on cells expressing α3β2γ2 receptors (10 inhibiting and 22

increasing the fluorescence quench). A third scan was conducted with the remaining active

test compounds that were again tenfold diluted (final concentrations of compounds per well

were between 503 and 593 nM). After that there were 47 active compounds on α1β2ε re-

ceptors (41 inhibiting and 6 enhancing the receptor activity) and seven on α3β2γ2 receptors

(one showed an inhibitory and 6 enhancing effects). Figures 5.5 and 5.6 represent results

of compounds that exhibited an effect on α1β2ε and α3β2γ2 GABAA receptors, respecti-

vely, at high concentrations, but both compounds failed to produce a similar effect at a lower

screening concentration. However the number of compounds which triggered an effect in

α1β2ε receptor-expressing cells during that last scan was still relatively high. Therefore a

further dilution of the remaining test compounds was carried out (concentrations from 50.3

to 59.3 nM) and tested. 17 compounds showed an effect on cells expressing α1β2ε receptors

whereas one compound enhanced receptor activity in cells expressing α3β2γ2 receptors.

The remaining 17 active compounds for α1β2ε receptors plus 2 further compounds that en-

hanced the fluorescence signal when using the next higher concentration were subject to

further investigation. Also applied in subsequent assays were the 7 compounds exhibiting an

effect on cells expressing α3β2γ2 receptors.

126


A B

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 5)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 4)


8.26 µM GABA + 0.5 % DMSO (n = 4)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.5: Results obtained for decreasing concentrations of a test compound in HEK293cells expressing α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA, 8.26 µM GABA alone(EC30), 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO or 8.26 µM GABAwith test compound TS-37. (A) 5.5 µM; (B) 550.4 nM. Concentrations are final concentrations perwell after injection of GABA. Data correspond to means ± SEM; n states the number of wells tested.Net fluorescence was normalised by setting the averaged baseline fluorescence to 100 % and fluores-cence changes (after adding solutions) were calculated as percentage of the baseline fluorescence.

50

60

70

80

90

100

110

120

130

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

130

5 10 15 20 25 30 35BF

100 µM GABA (n = 4)



Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

Figure 5.6: Results obtained for decreasing concentrations of a test compound in HEK293cells expressing α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after addition of 100 µM GABA, 2.45 µM GABA alone(EC30), 2.45 µM GABA and 1 mM picrotoxin, 2.45 µM GABA and 0.5 % DMSO or 2.45 µM GABAwith test compound TS-41. (A) 5.58 µM concentration of test compound, (B) 558.2 nM. Concen-trations are final concentrations per well after injection of solutions. Data correspond to means ±SEM; n states the number of wells tested. Net fluorescence was normalised by setting the averagedbaseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculated aspercentage of the baseline fluorescence.

127


5.2.2 Investigation of ’hit’ compounds on α1β2ε receptors

So far, test compounds were examined as singlets using only few controls (vehicle-only

controls and controls testing receptor functions). In order to rule out false positive results,

the remaining chemical compounds were examined in more detail using additional controls.

Each compound was tested in triplicates and on at least three different plates. Compounds

were applied at the lowest effective concentrations, determined before, to HEK293 cells ex-

pressing YFP-H148Q/I152L and the respective GABAA receptor subtype. NaI test solution

containing GABA at an EC30 concentration was added subsequently. Compounds were also

tested in the absence of GABA, injecting NaI test solution only. This would indicate pos-

sible GABA-independent effects of a test compound. Finally, compounds were applied at

high concentrations to HEK293 cells expressing mutant YFP-H148Q/I152L only. If fluores-

cence changes would be seen in these cells, it would indicate that a test compound acts on

endogenously expressed proteins of the cells rather than the specific GABAA receptor sub-

type. The latter two controls were more effective for compounds enhancing receptor activity.

Neither an agonist-independent effect nor an effect on endogenously expressed ion channels

could be seen from compounds that inhibit the receptors activity. However controls were

performed on all compounds in order to identify any possible irregularities.

Figures 5.7 to 5.25 display the results obtained for test compounds that showed an effect on

α1β2ε GABAA receptor activation during the initial screening assays in HEK293 cells, des-

cribed above. At a concentration of 54.58 nM, test compound TS-1 had shown an inhibiting

effect on the cells expressing the receptor subtype (Figure 5.7 A). However, after repeating

the screening assay using the compound in triplicates, the effect was less distinctive (but

significant) than before (Figure 5.7 B). Controls with test compounds only, in the absence of

GABA, or using cells expressing YFP-H148Q/I152L only, showed no YFP quench, which

was expected (see above; Figure 5.7 C, D).

128


70

80

90

100

110

5 10 15 20 25 30 35BF


8.26 µM GABA + 0.5 % DMSO (n = 4)

8.26 µM GABA (n = 6)8.26 µM GABA + 1 mM PTX (n = 3)

100 µM GABA (n = 3)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

80

90

100

110

5 10 15 20 25 30 35BF

100 µM GABA (n = 9)



Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)

BF

NaI (n = 16)NaI + compound TS-1 (n = 3)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

C D

Figure 5.7: Fluorescence signal change after adding test compound TS-1 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-1 (54.58 nM final concentration; initial screening assay). (B) The assay was repeated atleast three times using different plates and batches of cells. Effect of compound TS-1 was significantlydifferent to GABA-control response (p < 0.01). (C) Data from no-GABA control plates. NaI testsolution only was added to wells containing compound TS-1 (54.58 nM final concentration). (D)Effect of 8.26 µM GABA and 54.58 µM compound TS-1 in cells expressing YFP-H148Q/I152L only.100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 %DMSO and in (C) NaI test solution alone served as respective controls. Data correspond to means ±SEM; n states the number of wells measured. Net fluorescence of values was normalised by settingthe averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions) werecalculated as percentage of the baseline fluorescence.

129


Test compounds TS-2 as well as TS-3 inhibited GABA-control responses at concentrations

of 55.32 nM and 55.66 nM, respectively, during the initial screening assay. This could be

confirmed when repeating the assay using the compounds in triplicates (Figures 5.8 A, B

and 5.9 A, B). No effect could be seen in the absence of GABA or on cells expressing YFP-

H148Q/I152L only, which was expected (Figures 5.8 C, D and 5.9 C, D).

70

80

90

100

110

5 10 15 20 25 30 35


BF

8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BFTime [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)

NaI + compound TS-2 (n = 3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

Figure 5.8: Fluorescence signal change after adding test compound TS-2 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-2 (55.32 nM final concentration) in initial screening assay. (B) The assay was repeatedon at least three different plates. Effect of compound TS-2 was significantly different to GABA-control response (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was addedto wells containing compound TS-2 (55.32 nM final concentration). (D) Effect of 8.26 µM GABAand 55.32 µM compound TS-2 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

130


80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35


BF

8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.9: Fluorescence signal change after adding test compound TS-3 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-3 (55.66 nM final concentration) in initial screening assay. (B) The assay was repea-ted on three different plates. Effect of compound TS-3 was significantly different to GABA-controlresponse (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-3 (55.66 nM final concentration). (D) Effect of 8.26 µM GABA and55.66 µM compound TS-3 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as respective controls. Data correspond to means ± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

131


Test compound TS-4 inhibited the GABA-control response during the preliminary screening

assay (Figure 5.10 A). The assay was repeated using the compound in triplicates. The pre-

viously observed effect was less prominent, but still significant (Figure 5.10 B). Controls

with test compounds only, in the absence of GABA, or using cells expressing YFP-H148Q/

I152L only, showed no YFP quench (see above; Figure 5.10 C, D).

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

70

80

90

100

110

5 10 15 20 25 30 35

8.26 µM GABA + compound TS-4 (n = 1)8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.10: Fluorescence signal change after adding test compound TS-4 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-4 (54 nM final concentrations) in initial screening assay. (B) The assay was repeatedin triplicates on three different plates. Effect of compound TS-4 was significantly different to GABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solution only was addedto wells containing compound TS-4 (54 nM final concentration). (D) Effect of 8.26 µM GABA and54.01 µM compound TS-4 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

132


A similar result was obtained for compound TS-5. It inhibited the GABA-control response

during the preliminary screening assay (Figure 5.11 A), however, after repeating the assay

three times, the inhibition of the GABA-control response was less distinctive, but still signi-

ficant (Figure 5.11 B). Controls with test compounds only, in the absence of GABA, or using

cells expressing YFP-H148Q/I152L only, showed no YFP quench, which was expected (see

above; Figure 5.11 C, D).

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

70

80

90

100

110

5 10 15 20 25 30 35


BF

8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.11: Fluorescence signal change after adding test compound TS-5 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-5 (55.51 nM final concentration; initial screening assay). (B) The assay was repeatedthree times on different plates. Effect of compound TS-5 was significantly different to GABA-controlresponse (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-5 (55.51 nM final concentration). (D) Effect of 8.26 µM GABA and55.51 µM compound TS-5 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

133


The results obtained for compounds TS-6 and TS-7 also indicated the inhibition of the

GABA-control response in the first test (Figures 5.12 A and 5.13 A). These results could

be confirmed after repeating the assay, showing that the compounds triggered an effect si-

gnificantly different to the GABA-control responses (Figures 5.12 B and 5.13 B; C and D

show respective controls used).

80

90

100

110

120

130

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)NaI + compound TS-6 (n=3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35

8.26 µM GABA + compound TS-6 (n =1)

8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.12: Fluorescence signal change after adding test compound TS-6 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA in NaI test solution towells containing test compound TS-6 (55.02 nM final concentration; initial screening assay). (B)The assay was repeated three times on different plates. Effect of compound TS-6 was significantlydifferent to GABA-control response (p < 0.01). (C) Data from no-GABA control plates. NaI testsolution only was added to the wells containing compound TS-6 (55.02 nM final concentration). (D)Effect of 8.26 µM GABA and 55.02 µM compound TS-6 on cells expressing YFP-H148Q/I152Lonly. 100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and0.5 % DMSO and in (C) NaI test solution alone served as controls. Data correspond to means ±SEM; n states the number of wells measured. Net fluorescence of values was normalised by settingthe averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions) werecalculated as percentage of the baseline fluorescence.

134


A B

C D

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)NaI + compoundTS-7 (n = 3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35



100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.13: Fluorescence signal change after adding test compound TS-7 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-7 (54.42 nM final concentration; initial screening assay). (B) The assay was repeatedthree times on different plates using distinct batches of transfected cells. The effect of compoundTS-7 was significantly different to GABA-control response (p < 0.05). (C) Data from no-GABAcontrol plates. NaI test solution only was added to wells containing compound TS-7 (54.42 nM finalconcentration). (D) Effect of 8.26 µM GABA and 54.42 µM compound TS-7 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µMGABA and 0.5 % DMSO and in (C) NaI test solution alone served as controls. Data correspond tomeans± SEM; n states the number of wells measured. Net fluorescence of values was normalised bysetting the averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions)were calculated as percentage of the baseline fluorescence.

135


In contrast to the above compounds, test compounds TS-8 and TS-9 inhibited GABA-control

responses during the preliminary screening assay as well as when repeating the assay using

the compounds in triplicates on each plate (Figure 5.14 A, B and 5.15 A, B). No effects were

observed in the absence of GABA or in cells expressing YFP-H148Q/I152L only (Figures

5.14 C, D and 5.15 C, D).

A B

C D

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35

8.26 µM GABA +compound TS-8 (n = 1)

8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.14: Fluorescence signal change after adding test compound TS-8 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-8 (54.39 nM final concentration; initial screening assay). (B) The assay was repeatedthree times on different plates. Effect of compound TS-8 was significantly different to GABA-controlresponse (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-8 (54.39 nM final concentration). (D) Effect of 8.26 µM GABA and54.39 µM compound TS-8 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

136


A B

C D

70

80

90

100

110

5 10 15 20 25 30 35



100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)8.26 µM GABA +compound TS-9(n = 9)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)NaI + compound TS-9(n = 3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.15: Fluorescence signal change after adding test compound TS-9 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-9 (55.77 nM final concentration; initial screening assay). (B) Screening assay wasrepeated three times on different plates. Effect of compound TS-9 was significantly different toGABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solution onlywas added to wells containing compound TS-9 (55.77 nM final concentration). (D) Effect of 8.26 µMGABA and 55.77 µM compound TS-9 on cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in(C) NaI test solution alone served as respective controls. Data correspond to means ± SEM; n statesthe number of wells measured. Net fluorescence of values was normalised by setting the averagedbaseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculated aspercentage of the baseline fluorescence.

137


Compounds TS-10 and TS-11 also inhibited GABA-control responses during the preliminary

screen (Figures 5.16 A and 5.17 A), but the effects were less distinctive after repeating the

assays three times (Figures 5.16 B and 5.17 B). Compound TS-10 exhibited a fluorescence

quench of about 20 % in cells expressing YFP-H148Q/I152L only that was not seen in cells

expressing GABAA receptors, indicating iodide entrance into the cells (Figure 5.16 D).

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.16: Fluorescence signal change after adding test compound TS-10 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-10 (52.63 nM final concentration; initial screening assay). (B) The assay was repeatedthree times on different plates. The effect of the compound was significantly different to the GABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solution only was addedto the wells containing compound TS-10 (52.63 nM final concentration). (D) Effect of 8.26 µMGABA and 52.63 µM compound TS-10 in cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in(C) NaI test solution alone served as respective controls. Data correspond to means ± SEM; n statesthe number of wells measured. Net fluorescence of values was normalised by setting the averagedbaseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculated aspercentage of the baseline fluorescence.

138


In contrast, test compound TS-11 showed an increased fluorescence signal in cells expressing

YFP-H148Q/I152L only (Figure 5.17 D).

A B

C D

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

150

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)8.26 µM GABA +compound TS-11 (n = 6)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.17: Fluorescence signal change after adding test compound TS-11 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containing com-pound TS-11 (56.08 nM final concentrations) in initial screening assay. (B) The assay was repeatedon three different plates. The effect of the compound was significantly different to the GABA-controlresponse (p < 0.05). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-11 (56.08 nM final concentration). (D) Effect of 8.26 µM GABA and56.08 µM compound TS-11 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in (C) NaItest solution alone served as controls. Data correspond to means± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

139


Test compound TS-12 inhibited receptor activity in cells expressing α1β2ε receptors. Re-

peating the assay confirmed this result (Figure 5.18 A and B). NaI test solution control and

controls with cells lacking the receptors showed no effect upon the presence of the test com-

pound (Figure 5.18 C, D).

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.18: Fluorescence signal change after adding test compound TS-12 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-12 (54.2 nM final concentration; initial screening assay). (B) The assay was repeatedthree times on different plates. The effect of the compound was significantly different to the GABA-control response (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only wasadded to the wells containing compound TS-12 (54.2 nM final concentration). (D) Effect of 8.26 µMGABA and 54.2 µM compound TS-12 in cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in(C) NaI test solution alone served as respective controls. Data correspond to means ± SEM; n statesthe number of wells measured. Net fluorescence of values was normalised by setting the averagedbaseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculated aspercentage of the baseline fluorescence.

140


Results for test compounds TS-13 and TS-14 are depicted in Figures 5.19 and 5.20. The as-

say results indicated that the compounds inhibited GABA-control responses in cells expres-

sing α1β2ε GABAA receptors. Controls performed in cells expressing mutant YFP-H148Q/

I152L only showed increased fluorescence signals compared to the other controls applied

(Figures 5.19 and 5.20 D).

A B

C D

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

150

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)

8.26 µM GABA +compound TS-13 (n = 6)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.19: Fluorescence signal change after adding test compound TS-13 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containing com-pound TS-13 (55.34 nM final concentrations) in initial screening assay. (B) The assay was repeatedon three different plates. The effect of compound TS-13 was significantly different to the GABA-control response (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was addedto wells containing compound TS-13 (55.34 nM final concentrations). (D) Effect of 8.26 µM GABAand 55.34 µM compound TS-13 in cells expressing mutant YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in(C) NaI test solution alone served as controls. Data correspond to means± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

141


A B

C D

80

90

100

110

120

130

140

150

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)

NaI + compound TS-14(n = 3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.20: Fluorescence signal change after adding test compound TS-14 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containing com-pound TS-14 (55.03 nM final concentration) in initial screening assay. (B) The assay was repeatedon at least three different plates. The effect of compound TS-14 was significantly different to GABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solution only wasadded to wells containing compound TS-14 (55.03 nM final concentration). (D) Effect of 8.26 µMGABA and 55.03 µM compound TS-14 in cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in(C) NaI test solution alone served as controls. Data correspond to means± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

142


Test compound TS-15 inhibited the effect of GABA at EC30 concentrations in HEK293 cells

expressing α1β2ε receptors (Figure 5.21 A, B). However, when the test compound was ap-

plied to cells expressing YFP-H148Q/I152L only, the fluorescence signal decreased slightly

to 96.56 ± 1.18 % compared to the baseline fluorescence, whereas fluorescence signals of

all other controls were higher (Figure 5.21 D).

A B

C D

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35



100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.21: Fluorescence signal change after adding test compound TS-15 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-15 (54.97 nM final concentration; initial screening assay). (B) Assay was repeatedthree times on different plates. The effect of compound TS-15 was significantly different to theGABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solution onlywas added to wells containing compound TS-15 (54.97 nM final concentration). (D) Effect of 8.26 µMGABA and 54.97 µM compound TS-15 on cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in(C) NaI test solution alone served as respective controls. Data correspond to means ± SEM; n statesthe number of wells measured. Net fluorescence of values was normalised by setting the averagedbaseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculated aspercentage of the baseline fluorescence.

143


The majority of chemical compounds tested that showed an effect on cells expressing α1β2ε

GABAA receptors inhibited the fluorescence quench, and there were only a few test com-

pounds enhancing receptor activity. Therefore, fluorescence quench increasing compounds

that were identified at the second lowest concentration range tested (between 503 and 593 nM)

were also examined by repeating the assays with respective controls. Test compound TS-16

triggered a fluorescence quench that was larger than the quench obtained when cells were

exposed to 100 µM GABA. This could be confirmed by repeating the assay several times

in triplicates (Figure 5.22 A, B). Applying the compound with NaI test solution only also

indicated a small effect (Figure 5.22 C). In both cases, the onset of the responses was rela-

tively slow compared to the GABA-control responses. Cells lacking the GABAA receptors

showed no change in fluorescence signal when exposed to the compound, indicating that the

response was GABAA receptor-dependent.

144


A B

C D

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)NaI + compound TS-16 (n = 11)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

60

70

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 3)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.22: Fluorescence signal change after adding test compound TS-16 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-16 (539.4 nM final concentration) in initial screening assay. (B) The assay was repea-ted several times using different plates. The effect of compound TS-16 was significantly differentto GABA-control response (p < 0.05). (C) Data from no-GABA control plates. NaI test solutiononly was added to the wells containing compound TS-16 (539.4 nM final concentration). (D) Effectof 8.26 µM GABA and 53.94 µM compound TS-16 on cells expressing YFP-H148Q/I152L only.100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 %DMSO and in (C) NaI test solution alone served as respective controls. Data correspond to means ±SEM; n states the number of wells measured. Net fluorescence of values was normalised by settingthe averaged baseline fluorescence to 100 % and fluorescence changes (after adding solutions) werecalculated as percentage of the baseline fluorescence.

145


In contrast, the increased fluorescence quench triggered by compound TS-17 that was ob-

served during the initial screening assay could not be confirmed when repeating the assay;

no significant change could be seen compared to the EC30 GABA-control response (Figure

5.23). Similar results were obtained for compounds TS-18 and TS-19. The compounds in-

hibited the fluorescence change evoked by GABA during the initial screens, but in further

replicates no significant effect could be observed (Figures 5.24 and 5.25).

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

A B

C D

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 3)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.23: Fluorescence signal change after adding test compound TS-17 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containing com-pound TS-17 (570 nM final concentration; initial screening assay). (B) The assay was repeated threetimes on different plates. (C) Data from no-GABA control plates. NaI test solution only was addedto the wells containing compound TS-17 (570 nM final concentration). (D) Effect of 8.26 µM GABAand 57 µM compound TS-17 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

146


A B

C D

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)NaI + compound TS-18 (n = 3)

BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

150

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO(n = 23)8.26 µM GABA +compound TS-18(n = 9)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.24: Fluorescence signal change after adding test compound TS-18 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containingcompound TS-18 (54.02 nM final concentration) in initial screening assay. (B) The screening assaywas repeated on different plates. (C) Data from no-GABA control plates. NaI test solution only wasadded to wells containing compound TS-18 (54.02 nM final concentration). (D) Effect of 8.26 µMGABA and 54.02 µM compound TS-18 on cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in(C) NaI test solution alone served as controls. Data correspond to means± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

147


A B

C D

80

90

100

110

5 10 15 20 25 30 35

100 µM GABA (n = 9)


8.26 µM GABA + 0.5 % DMSO (n = 18)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

5 10 15 20 25 30 35


8.26 µM GABA + 0.5 % DMSO (n = 4)


100 µM GABA (n = 3)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 9)


NaI + 0.5 % DMSO (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.25: Fluorescence signal change after adding test compound TS-19 to HEK293 cellsexpressing YFP-H148Q/I152L and α1β2ε GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 8.26 µM GABA to wells containing com-pound TS-19 (55.33 nM final concentrations) in initial screening assay. (B) The assay was repeatedon three different plates. (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-19 (55.33 nM final concentration). (D) Effect of 8.26 µM GABA and55.33 µM compound TS-19 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in (C) NaI testsolution alone served as respective controls. Data correspond to means ± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

148


5.2.3 Investigation of ’hit’ compounds on α3β2γ2 receptors

Test compounds were also applied to HEK293 cells expressing mutant YFP-H148Q/I152L

and α3β2γ2 GABAA receptors. In contrast to cells expressing α1β2ε GABAA receptors, the

majority of compounds applied to cells expressing α3β2γ2 receptors enhanced the YFP fluo-

rescence quench during the preliminary screening tests. Furthermore, the number of active

compounds was less and only seven were left after preliminary screenings at a concentration

range of 503 to 593 nM. These were tested further by repeating the assays in several wells

per plate and also using further controls, as described above for Section 5.2.2.

Test compounds TS-10, TS-20 and TS-21 elicited enhanced quenching of the mutant YFP

compared to the GABA-control response (EC30) in cells expressing α3β2γ2 receptors during

the preliminary screens. This was confirmed when repeating the assays using triplicates per

plate (Figures 5.26 A, B, 5.27 A, B and 5.28 A, B). The compounds also caused a fluores-

cence quench when applied in the absence of GABA. However, when test compounds were

added to control cells expressing YFP-H148Q/I152L only, a fluorescence quench could also

be observed, indicating the interference of some endogenously expressed proteins (Figures

5.26 C, D, 5.27 C, D and 5.28 C, D).

Test compound TS-22 inhibited the GABA-elicited fluorescence quench during the prelimi-

nary screens (Figure 5.29 A). Repeating the test confirmed this result, although the effect

was less prominent compared to before (Figure 5.29 B). The fluorescence signal was similar

to that of the controls when applying compound TS-22 without GABA to the cells. However,

performing the assay on cells expressing mutant YFP-H148Q/I152L only, the fluorescence

signal was higher than that obtained from the respective controls (100 µM GABA, 8.26 µM

GABA, 8.26 µM GABA and 1 mM picrotoxin and 8.26 µM GABA with 0.5 % DMSO;

Figure 5.29 C, D).

149


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35


2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35


NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)

NaI (n = 16)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.26: Fluorescence signal change after adding test compound TS-10 to HEK293 cellsexpressing mutant YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing526.3 nM compound TS-10 (final concentrations) in initial screening assay. (B) The assay was re-peated in triplicates on different plates. The effect of compound TS-10 was significantly differentto GABA-control response (p < 0.001). (C) Data from no-GABA control plates. NaI test solutiononly was added to the wells containing compound TS-10 (526.3 nM final concentration). (D) Effectof 8.26 µM GABA and 52.63 µM compound TS-10 on cells expressing YFP-H148Q/I152L only.100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 %DMSO and in (C) NaI test solution alone served as controls. Data correspond to means ± SEM; nstates the number of wells measured. Net fluorescence of values was normalised by setting the avera-ged baseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculatedas percentage of the baseline fluorescence.

150


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)

NaI + compound TS-20 (n = 9)NaI (n = 16)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.27: Fluorescence signal change after adding test compound TS-20 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing543.1 nM compound TS-20 (final concentrations) in initial screening assay. (B) The assay was repea-ted on several different plates. The effect of compound TS-20 was significantly different to GABA-control response (p < 0.001). (C) Data from no-GABA control plates. NaI test solution only wasadded to wells containing compound TS-20 (543.1 nM final concentration). (D) Effect of 8.26 µMGABA and 54.31 µM compound TS-20 on cells expressing YFP-H148Q/I152L only. 100 µM GABA,8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in(C) NaI test solution alone served as controls. Data correspond to means± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

151


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

40

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)


2.45 µM GABA + 0.5 % DMSO (n = 4)2.45 µM GABA + compound TS-21(n = 1)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.28: Fluorescence signal change after adding test compound TS-21 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing548.9 nM compound TS-21 (final concentrations) in initial screening assay. (B) Screening assay wasrepeated on different plates. The effect of compound TS-21 was significantly different to GABA-control response (p < 0.001). (C) Data from no-GABA control plates. NaI test solution only wasadded to wells containing compound TS-21 (548.9 nM final concentration). (D) Effect of 8.26 µMGABA and 54.89 µM compound TS-21 was investigated on cells expressing YFP-H148Q/I152L only.100 µM GABA, 8.26 µM GABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 %DMSO and in (C) NaI test solution alone served as controls. Data correspond to means ± SEM; nstates the number of wells measured. Net fluorescence of values was normalised by setting the avera-ged baseline fluorescence to 100 % and fluorescence changes (after adding solutions) were calculatedas percentage of the baseline fluorescence.

152


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

150

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)

8.26 µM GABA +compound TS-22(n = 9)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.29: Fluorescence signal change after adding test compound TS-22 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing545.8 nM compound TS-22 (final concentrations) in initial screening assay. (B) The assay was repea-ted on different plates. The effect of compound TS-22 was significantly different to GABA-controlresponse (p < 0.01). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-22 (545.8 nM final concentration). (D) Effect of 8.26 µM GABA and54.58 µM compound TS-22 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in (C) NaItest solution alone served as controls. Data correspond to means± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

153


Compounds TS-23 and TS-16 had increased the fluorescence quench compared to the GABA-

control response in cells expressing α3β2γ2 GABAA receptors. Results from the former

could be confirmed while repeating the assay several times on different plates. The fluores-

cence quench obtained was 49.19 ± 3.27 % and therefore slightly higher than the quench

obtained for 100 µM GABA (44.15 ± 1.22 %; Figure 5.30 A, B). The compound also trig-

gered a small fluorescence quench when applied without GABA. No effect was seen in cells

lacking α3β2γ2 GABAA receptors, indicating that the observed YFP quench is GABAA re-

ceptor specific (Figure 5.30 C, D). Similar results were obtained for the latter compound,

TS-16. However, the fluorescence quench was weaker (38.15 ± 3.65 %); no effect was

seen when the compound was applied without GABA or to cells lacking the receptor (Figure

5.31).

The positive screening result obtained in the initial screening tests using test compound TS-

24 turned out to be false positive. The fluorescence quench could not be confirmed when

repeating the assays (Figure 5.32).

The number of test compounds enhancing or inhibiting GABAA receptor activity determi-

ned by the preliminary screening tests (Section 5.2.1) could be further reduced by subtracting

false-positive results and those compounds that triggered an effect on cells lacking the re-

ceptors when retesting the compounds using triplicates per plate (see above). The remaining

compounds exhibiting GABAA receptor-specific activity could now be further investigated.

There were seven compounds left exhibiting a distinctive effect on α1β2ε receptors and three

compounds that showed an effect on α3β2γ2 GABAA receptors.

154


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)


BF

NaI (n = 16)

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)


8.26 µM GABA + 0.5 % DMSO (n = 23)

8.26 µM GABA +compound TS-23(n = 9)

BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.30: Fluorescence signal change after adding test compound TS-23 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing546.6 nM compound TS-23 (final concentrations) in initial screening assay. (B) The assay was repea-ted on different plates. The effect of compound TS-23 was significantly different to GABA-controlresponse (p < 0.001). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-23 (546.6 nM final concentration). (D) Effect of 8.26 µM GABA and54.66 µM compound TS-23 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in (C) NaI testsolution alone served as respective controls. Data correspond to means ± SEM; n states the numberof wells measured. Net fluorescence of values was normalised by setting the averaged baseline fluo-rescence to 100 % and fluorescence changes (after adding solutions) were calculated as percentage ofthe baseline fluorescence.

155


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 3)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.31: Fluorescence signal change after adding test compound TS-16 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing539.4 nM compound TS-16 (final concentrations) in initial screening assay. (B) The assay was repea-ted on different plates. The effect of compound TS-16 was significantly different to GABA-controlresponse (p < 0.05). (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-16 (539.4 nM final concentration). (D) Effect of 8.26 µM GABA and53.94 µM compound TS-16 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA and 0.5 % DMSO and in (C) NaI testsolution alone served as controls. Data correspond to means ± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

156


A B

C D

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

2.45 µM GABA + 0.5 % DMSO (n = 13)


100 µM GABA (n = 12)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

NaI + 0.5 % DMSO (n = 16)


100 µM GABA (n = 15)


BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

70

80

90

100

110

120

130

140

5 10 15 20 25 30 35

100 µM GABA (n = 12)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

50

60

70

80

90

100

110

120

5 10 15 20 25 30 35

100 µM GABA (n = 4)



BF

Time [sec]

% N

et

flu

ore

sce

nce

sig

nal

Figure 5.32: Fluorescence signal change after adding test compound TS-24 to HEK293 cellsexpressing YFP-H148Q/I152L and α3β2γ2 GABAA receptors.Fluorescence was measured for 30 seconds after adding (A) 2.45 µM GABA to wells containing553.6 nM compound TS-24 (final concentrations) in initial screening assay. (B) The assay was repea-ted on different plates. (C) Data from no-GABA control plates. NaI test solution only was added towells containing compound TS-24 (553.6 nM final concentration). (D) Effect of 8.26 µM GABA and55.36 µM compound TS-24 on cells expressing YFP-H148Q/I152L only. 100 µM GABA, 8.26 µMGABA, 8.26 µM GABA and 1 mM picrotoxin, 8.26 µM GABA with 0.5 % DMSO and in (C) NaItest solution alone served as controls. Data correspond to means± SEM; n states the number of wellsmeasured. Net fluorescence of values was normalised by setting the averaged baseline fluorescence to100 % and fluorescence changes (after adding solutions) were calculated as percentage of the baselinefluorescence.

157


5.2.4 Investigation of ’hit’ compounds in Xenopus laevis oocytes

Next, a selection of the remaining test compounds was examined in more detail using the

two-electrode voltage-clamp technique by expressing α1β2ε or α3β2γ2 GABAA receptors

in Xenopus laevis oocytes. Test compound TS-3 was selected due to its significant inhibitory

effect on cells expressing the α1β2ε receptor and the fact that control experiments confir-

med the results (Figure 5.9). Applying increasing concentrations of compound TS-3 with

GABA (EC30) to Xenopus oocytes expressing α1β2ε receptors did not show significant dif-

ferences to currents obtained from GABA-control responses (n = 4; Figure 5.33 A, B) and

the inhibiting effect of the compound could not be confirmed.

The mutant YFP-based screening assays indicated that test compound TS-16 enhanced both

α1β2ε and α3β2γ2 GABAA receptors expressed in HEK293 cells (Figures 5.22 and 5.31).

It was the only test compound identified in the screening assay that enhanced the activity

of α1β2ε receptors and, therefore, further investigated. The compound’s enhancing effect

could be confirmed in oocytes expressing α3β2γ2 receptors (Figure 5.34 A, C). The GABA-

control response increased by 65.8 % when applying 10 or 100 µM compound TS-16 together

with GABA (EC20) to the oocytes. The compounds EC50 was 3.67 µM (n = 7). However, in

oocytes expressing α1β2ε receptors, no clear concentration-dependent characteristics could

be detected when applying the compound (n = 4; Figure 5.34 A, B). High concentrations of

the test compound actually triggered an “outward” current in two of four oocytes. Overall

the current traces obtained throughout for α1β2ε receptors expressed in Xenopus oocytes

were very small compared to traces from other GABAA receptor subtypes.

Since the results so far obtained for α1β2ε GABAA receptors were inexplicit, the final com-

pound tested in oocytes was compound TS-23. Compound TS-23 enhanced GABA-evoked

YFP quench in the screening assay setup using HEK293 cells expressing α3β2γ2 recep-

tors (Figure 5.30). The results obtained from oocytes confirmed the potentiation of GABA-

mediated responses by the compound. At concentrations of 10 µM or higher, the response

was about 60 % enhanced compared to the GABA-control response. The EC50 of the com-

pound was 1.42 µM (n = 5; Figure 5.35 A, B).

158


-9 -8 -7 -6 -5 -40

25

50

75

100

125

150

α1β2ε

log [Compound TS-3, M]

% G

AB

A c

on

tro

l re

spo

nse

10 µM1 µMGABA

20 nA

50 sec

3 µM 30 µM 100 µM

A

B

Figure 5.33: Results obtained from oocytes expressing α1β2ε receptors after exposure totest compound TS-3 and GABA at EC30 concentrations.(A) Compound TS-3 concentration response analysis using a non-linear regression fit with variableslopes as described in Section 2.3.5 (n = 4). Data points correspond to means ± SEM; n states thenumber of oocytes examined. (B) Representative current traces evoked by GABA at EC30 concen-trations alone or in combination with increasing concentrations of compound TS-3. Horizontal barsrepresent drug application.

159


-9 -8 -7 -6 -5 -40

25

50

75

100

125

150

175

α3β2γ2

α1β2ε


% G

AB

A c

on

tro

l re

spo

nse

100 sec

500 nA

GABA 1 nM 10 nM 100 nM 1 µM 10 µM

A

B

C

20 nA

50 sec

GABA 100 nM 10 µM1 µM10 nM

Figure 5.34: Results obtained from oocytes expressing α1β2ε or α3β2γ2 receptors afterexposure to test compound TS-16 and submaximal GABA concentrations.(A) Compound TS-16 concentration-response analysis using a non-linear regression fit with variableslopes as described in Section 2.3.5 (n = 4 for α1β2ε and n = 7 for α3β2γ2). Data points correspondto means ± SEM; n states the number of oocytes examined. (B) Representative current traces evokedby GABA at EC30 concentrations alone or in combination with increasing concentrations of com-pound TS-16 in oocytes expressing α1β2ε receptors. (C) Sample current traces triggered by GABAat EC20 concentrations alone or in combination with increasing concentrations of compound TS-16in oocytes expressing α3β2γ2 receptors. Horizontal bars represent drug application.

160


-9 -8 -7 -6 -5 -4

75

100

125

150

175

α3β2γ2


% G

AB

A c

on

tro

l re

spo

nse

100 sec

500 nA

GABA 1 nM 10 nM 100 nM 1 µM 10 µM

A

B

Figure 5.35: Results obtained from oocytes expressing α3β2γ2 receptors after exposure totest compound TS-23 and GABA at EC30 concentrations.(A) Compound TS-23 concentration-response analysis using a non-linear regression fit with variableslopes as described in Section 2.3.5 (n = 5). Data points correspond to means ± SEM; n states thenumber of oocytes examined. (B) Sample current traces triggered by GABA at EC30 concentrationsalone or in combination with increasing concentrations of compound TS-23 in oocytes expressingα3β2γ2 receptors. Horizontal bars represent drug application.

161


5.3 Discussion

After setting up and optimising a cell-based screening assay using the mutant YFP-H148Q/

I152L, 800 chemical compounds of the Maybridge HitFinder™ Collection were screened

against α1β2ε as well as α3β2γ2 GABAA receptors expressed in HEK293 cells using a

fluorescence plate reader. As mentioned in the previous chapter, the assay setup using cells

expressing α1β2ε receptors exhibited a very small signal range and Z’-factor and its suita-

bility to screen for compounds was uncertain. However, it was proposed that YFP quench-

inhibiting or -enhancing compounds could be identified visually and, thus, the assay format

was applied due to the lack of alternative methods. By screening compounds against both

receptor subtypes it was possible to compare compound effects and to identify compounds

that were specific to one of the receptor subtypes.

In order to elicit the desired biological action, it is necessary to apply a chemical compound

at a specific concentration (Inglese et al., 2006). Moreover, the right compound concentra-

tion for the preliminary screening is often chosen according to the capacity of the secondary

screening test or further investigations (Willumsen et al., 2003). In this project, compounds

were screened initially at relatively high concentrations with the aim to evaluate the hit rate.

The majority of the test compounds did show little or no effects. This was expected since in

unbiased libraries most compounds do not exhibit activity (Zhang et al., 1999). Nevertheless,

the hit rate was relatively high and numerous compounds seemed to exhibit effects by enhan-

cing or inhibiting YFP quenches on the two receptor subtypes. The question arose whether

the chosen test compound concentration was too high, and the fluorescence change seen in

many samples may have been due to “off-target” or toxic effects, which are sometimes ob-

served at high compound concentrations (Inglese et al., 2006). In a study performing HTS

on 5-HT6Rs, which are G protein-coupled receptors, chemical compounds at a concentration

of 10 µM were applied (Kim et al., 2008). However, since the nature of the performed assays

and investigated proteins was very different, compound concentrations cannot be compared

easily. Here, in order to identify the most potent compounds, their concentrations were gra-

dually decreased and retested until a manageable number of compounds was left for further

162


investigations.

When examining the preliminary screening results it is striking that the majority of com-

pounds influencing the fluorescence signal in cells expressing α1β2ε receptors are inhibiting

the GABA-evoked YFP quench. In contrast, most compounds increased the fluorescence

quench in cells expressing α3β2γ2 receptors. Furthermore, the compound concentrations

used in the later screens were relatively low, especially after the last dilution. However, since

all compounds tested at this concentration inhibited the YFP quench in α1β2ε receptors,

it is possible that some inhibited the open, spontaneously active receptor channels before

GABA in NaI solution was injected into the wells. The test compounds were added to 50 µl

NaCl bathing solution before NaI test solution was injected (50.3 to 59.3 nM was the final

concentration of compounds per well).

The preliminary screening tests were performed in singlets and the occurrence of false po-

sitive results among the selected compounds was expected, since they are very common in

single-dose, single-experiment assay setups (Hann & Oprea, 2004). There are many dif-

ferent reasons for false positive results to occur: it could be due to pipetting errors, cross-

contamination or other handling errors, also data handling and analysis. These could have

easily happened despite the great care that was taken due to the large amount of compounds

and subsequent data obtained that were all handled manually. Further reasons are possible

quenching artefacts or auto-fluorescence of compounds, variability of the transfected cells,

measurements on the edge of the plate, reagent differences or other effects triggered by a

compound independently from the GABAA receptor subtypes investigated (Gonzalez & Ma-

her, 2002; Gill et al., 2003; Terstappen, 2005; Inglese et al., 2007a; Sui & Wu, 2007; Janzen

& Popa-Burke, 2009).

Test compounds exhibiting positive effects during the preliminary screening tests were fur-

ther examined to eliminate possible false positive results. Thus, the assays were repeated

using triplicates as well as additional controls, such as cells lacking the GABAA receptors

and, to check for agonist-independent activity, applying the compounds without submaximal

GABA concentrations. Note that the latter control does not function with compounds that

163


inhibited GABA-control responses. By this means, several test compounds that previously

inhibited or enhanced GABA-elicited fluorescence quenches could be ruled out for further

testing. Of those, a number of compounds showed a YFP quench on cells lacking the recep-

tors. Hence, it was suggested that the effect seen after applying the compound was caused

via endogenously expressed proteins. The same might be the case where the fluorescence si-

gnal was higher compared to the other controls used in cells lacking receptors. Although it is

not clear, why the fluorescence signal would increase compared to the baseline fluorescence;

it could also be some assay artefact, since the fluorescence signal in no-GABA controls did

not show a comparable increase. Further compounds could be eliminated due to the fact that

the effect seen at first was less prominent or non existent when repeating the assay; the latter

may have happened due to pipetting or data handling errors.

After performing the mutant YFP-based assay, from 800 compounds screened, seven were

identified to be active on α1β2ε receptors and three on α3β2γ2 GABAA receptors. To do a

different type of assay is a frequently applied method to confirm the activity of a test com-

pound (Zheng et al., 2004). Here, the two-electrode voltage-clamp technique was used to

confirm and examine the compound’s activities in more detail. Due to time limitations, not

all of the identified compounds could be examined in oocytes. The first compound investi-

gated was compound TS-3. This compound showed no effects on cells expressing α3β2γ2

GABAA receptors, but inhibited the GABA-evoked fluorescence quench in cells expressing

α1β2ε receptors in the YFP-based assay. However, this effect could not be confirmed when

using the two-electrode voltage-clamp technique and Xenopus oocytes. The problem was

that currents from oocytes expressing α1β2ε receptors were very small, and therefore it was

difficult to distinguish differences between currents elicited by submaximal concentrations

of GABA and those evoked by the compound together with GABA. Additionally, some of

the valves of the perfusion system were stuck and opened manually. It is therefore likely that

the data obtained do not reflect the effect of the test compound but rather unsteadiness of the

perfusion of the compound. The next compound examined in oocytes was compound TS-16,

which increased the YFP quench on both receptor subtypes tested in HEK293 cells using the

YFP-based screening assay. Again, the results obtained for the ε-containing receptor sub-

164


type were ambiguous. Interestingly, instead of increasing GABA-evoked currents, at higher

concentrations an “outward” current could be seen in two of four oocytes tested that was not

seen with lower concentrations or GABA alone. This may indicate that, at higher concentra-

tions, the compound blocked the spontaneously open channels. Tests in oocytes expressing

α3β2γ2 receptors confirmed the results of the YFP-based assay; the compound enhanced

GABA-evoked currents in a concentration-dependent manner. Due to the difficulties with

the oocytes expressing the ε-containing receptors, the last investigated compound was com-

pound TS-23, which enhanced the YFP quench in HEK293 cells expressing the α3β2γ2

receptor. The activity of the compound on this receptor subtype could also be confirmed in

oocytes.

In summary, performing two-electrode voltage-clamp experiments in oocytes expressing

α3β2γ2 GABAA receptors could confirm the results obtained from the YFP-based scree-

ning assay in HEK293 cells. To what extend results are comparable is not certain, since the

compounds were only tested at a single concentration using the cell-based screening assay,

except when testing GABA-concentration responses (see Section 4.3.3, Chapter 4). Never-

theless, from the results obtained here, it can be concluded that the screening assay is a

suitable assay format to screen chemical compound libraries against α3β2γ2 receptors. The

EC50 values obtained for the two hit compounds were in the low micromolar range. Pos-

sible future work may lead to generation and screening of compound analogues which may

enhance the potencies of the compounds (see below).

Results obtained for α1β2ε receptors were ambiguous when using the two-electrode voltage-

clamp technique and no subtype-selective compound could be identified yet. This may partly

be a reflection on the cell-based screening assay that exhibited a small signal range and, thus,

low robustness. Nevertheless, it was unexpected that the obtained results could not be confir-

med performing the two-electrode voltage-clamp technique in Xenopus oocytes. It may be

worth trying the patch-clamp method in the future and test the compounds on receptors ex-

pressed in HEK293 cells, rather than oocytes, since the compounds were identified using

these cells. Furthermore, oocytes are less sensitive during pharmacological investigations

compared to mammalian cells (Xu et al., 2001). In the future, the ability to identify hit com-

165


pounds could also be improved, if the assays signal range and, thus, quality (Z’-factor) could

be further enhanced. One possibility to improve the assay’s quality could be by manipulating

the ε subunit in a way that the continuous activity could be eliminated without influencing

agonist or modulator binding properties, because amino acid changes would only be ge-

nerated in the pore lining of the channel. One would need to keep in mind that there are

modulators that bind to residues in the channels pore. However, those are not a priority in

the present study mainly seeking test compounds binding elsewhere on the receptor channel

(see Section 1.4, Chapter 1).

Originally, it was planned to verify a number of hit compounds and to generate a second-

generation library, consisting of structural analogues of the compounds exhibiting the stron-

gest effects during the preliminary screening. Afterwards, pharmacological characteristics

of the improved hit compounds could be examined using electrophysiological techniques. In

drug screening tests, the screening of hit analogues generated via parallel synthesis is com-

mon practice, because a compound’s properties towards the target protein can be enhanced

(van Niel et al., 2005; Kim et al., 2008; Wafford et al., 2009). Structure-activity relationships

generated by screening compound analogues would reveal which structural elements are cru-

cial for biological efficacy and would deliver first-generation tools for examining the in vivo

function of ε-subunit-containing receptors. In this study, hit compounds were tested im-

mediately after the preliminary screening in oocytes using the two-electrode voltage-clamp

technique due to time limitations of this project. In the future, due to the fact that it was

possible to screen for compounds modulating GABAA receptor activity using this assay for-

mat, further chemical compounds may be screened and analogues of hit compounds may be

subsequently generated to further evaluate and examine their properties.

166

Chapter 6

General Discussion

GABAA receptors are targeted by a plethora of different clinically important drugs. Howe-

ver most of them target the receptors non-selectively, hence adverse side effects often occur

(Buffett-Jerrott & Stewart, 2002; Möhler et al., 2002; Bateson, 2004; Agid et al., 2007). Ne-

vertheless, the structural diversity of the receptors offers great prospects for finding subunit

selective modulatory compounds that would target specific receptor subtypes.

Different receptor subunits confer distinct pharmacological and electrophysiological proper-

ties to GABAA receptor subtypes (Sieghart & Sperk, 2002). The ε subunit has only been

investigated in few studies and not much is known about its physiological role in vivo. Pre-

vious studies that investigated pharmacological properties of ε-containing receptors showed

that the subunit confers different properties to the receptor, compared to αβ or αβγ recep-

tors, however, no subunit-selective pharmacological profile could be established (Olsen &

Sieghart, 2009).

6.1 Pharmacological characterisation of the α3β2εGABAA

receptor

The pharmacological properties of α3β2εGABAA receptors were investigated using the two-

electrode voltage-clamp technique. This subunit combination had not been examined before.

Because the β2 subunit is the most widely distributed β subunit in the mammalian brain, it

167

Chapter 6. General Discussion

is likely that the subunits form functional channels in some areas of the brain. The data

obtained were similar to results obtained in previous recombinant pharmacological studies

investigating the ε subunit in combination with different α and β subunits (Davies et al.,

1997; Whiting et al., 1997; Thompson et al., 1998; Neelands et al., 1999; Maksay et al., 2003;

Ranna et al., 2006). It was shown that the inclusion of the ε subunit into a GABAA receptor

complex changes its pharmacological properties, for example α3β2ε receptors exhibited

high agonist sensitivity compared to α3β2 and α3β2γ2 receptors, but the potency to the

anaesthetic etomidate was less compared to α3β2 receptors. Nevertheless, no ε-selective

compounds are known yet. However, it is possible that there is a specific binding site located

between the α and ε subunits, probably similar to the benzodiazepine binding site (due to

sequence similarities between different subunits) and it was suggested that the ε subunit

represents an important future drug target (Whiting, 1999). It was further shown that the

ε subunit confers unusual properties to α3β2ε receptors, for example constitutive receptor

activity, which was also reported from other recombinant receptor subtypes containing the ε

subunit (Whiting et al., 1997; Neelands et al., 1999; Wagner et al., 2005). It is not known,

whether ε-containing receptors exhibit spontaneous activity in vivo and what implications

this would have on the receptors function or localisation.

Many GABAA receptor subunits exhibit splice versions, however they are usually restricted

to a single splice site, whereas the ε subunit exhibits multiple splice sites (Whiting et al.,

1997; Wilke et al., 1997; Erlitzki et al., 2000; Sinkkonen et al., 2000; Kasparov et al., 2001).

Many of these are expressed in the peripheral tissue, but there are also distinct splice versions

present in the brain and the question arises: why are there so many ε subunit splice variants?

Wilke et al. (1997) suggested that truncated transcripts may be expressed to down-regulate ε

subunit expression or they could have another, as yet unknown, function. The work conduc-

ted here showed that the εT subunit version expressed in brain (described by Kasparov et al.,

2001) is not incorporated into functional receptors when co-expressed with α3 and β2 subu-

nits in Xenopus oocytes. Similar results have also been obtained for other ε subunit splice

variants and the only functional variant known so far is the ε transcript of 3.5 kb found in

brain (Whiting et al., 1997; Davies et al., 2002).

168


Due to the fact that not much is known about the ε subunit and no specific pharmacology has

been identified yet, it was decided to search for ε-selective compounds. Previous studies have

screened compound libraries against specific α subunits or δ-subunit-containing receptors

(van Niel et al., 2005; Wafford et al., 2009), but no study has been conducted to search

for ε-subunit-selective modulatory compounds. Such a compound could be an important

pharmacological tool to further study the role of the promiscuous ε subunit and it may also

be a key step towards the development of future drugs targeted to ε-containing receptors,

for example to treat AD. Therefore, the next goal of the study was to set up a cell-based

screening assay to screen the ε-subunit-containing receptor subtype for selective modulatory

compounds.

6.2 Search for and setup of a cell-based screening assay

Electrophysiological recording techniques, such as the two-electrode voltage-clamp tech-

nique, are labour intensive with low throughput, and hence unsuitable for the screening of

compound libraries. Thus, a cell-based screening assay was sought. Several methods exist to

screen GABAA receptors (see Smith & Simpson, 2003; Gilbert et al., 2009b; Verkman & Ga-

lietta, 2009). However, only few were relatively easy to set up and did not require specialised

measuring devices, but could be performed using fluorescence or absorbance plate readers.

Of the three assay formats tested, the last attempted assay approach, applying a mutant YFP,

finally gave the desired results. The mutant YFP-H148Q/I152L acts as an iodide-ion indi-

cator protein and was expressed together with α3β2γ2 GABAA receptors in HEK293 cells.

The cells were plated into a 96-well microtitre plate and exposed to high concentrations

of the agonist GABA. This activated the expressed receptors facilitating iodide influx into

the cells that quenched the YFP, which could be measured using a fluorescence plate rea-

der. Fluorescence quench was prevented when the GABAA receptor blocker picrotoxin was

applied. Kruger et al. (2005) tested the assay format in GlyRs as well as α1β1 GABAA

receptors and obtained a similar fluorescence quench for the GABAA receptor subtype as

was obtained in this study using α3β2γ2 receptors. After it was shown that the assay format

was functional, cells expressing the α3β2ε receptor were tested, but no fluorescence quench

169


could be detected. Thereupon, several other receptor subtypes were tested, indicating that

the maximum fluorescence quench could vary widely, depending on the subunit composition

of the expressed receptor. It was proposed that these differences are due to different kine-

tic and gating properties of the receptor subtypes (that are conferred by different subunits).

Furthermore, differences in transfection efficiency also influenced the YFP quench (Gilbert

et al., 2009b). However, it was unclear why no fluorescence quench was detected in cells

expressing the α3β2ε-containing receptor. It was proposed that the constitutive activity of

the receptor subtype could have also influenced the result. In a final attempt it was exa-

mined whether the inclusion of the α1 subunit, instead of the α3 subunit, into the receptor

complex would change the assays outcome, and, surprisingly, a small fluorescence quench

could be detected. Although the YFP quench was less considerable than the one obtained

from α3β2γ2 receptors, the α1β2ε receptor subtype was subsequently used in the screening

assay.

The results of the mutant YFP screening assay setup once again demonstrated that the ε su-

bunit confers different properties to GABAA receptors compared to other receptor subunits.

During the two-electrode voltage-clamp studies it was noticeable that GABA-evoked cur-

rents were smaller in oocytes expressing α3β2ε receptors compared to oocytes expressing

α3β2 or α3β2γ2 receptors. The small signal range obtained in the mutant YFP screening

assay setup raised the question whether there is a relation between the two findings. The

spontaneous activity observed in recombinant ε-containing receptors influences the current

size and probably the signal window of the plate reader assay. However, this is likely not

the only reason. It could be that ε-subunit-containing receptors were less efficiently expres-

sed on the cell surface compared to other GABAA receptors. Gilbert et al. (2009b) showed

that cells expressing fewer receptors and YFP (due to different transfection methods) also

exhibited a smaller % fluorescence quench.

170


6.3 Screening of α1β2ε and α3β2γ2 GABAA receptors

After the assay format was optimised and its functionality demonstrated, a chemical com-

pound library was screened against both α1β2ε and α3β2γ2 GABAA receptors expressed in

HEK293 cells. The assay setup to screen the former subtype was not optimal for compound

screening due to a small signal dynamic range. However, it was possible to distinguish the

effect of compounds that inhibited or enhanced the fluorescence quench from the effect evo-

ked by GABA at EC30 concentrations. Screening data were compared with those obtained

from screening α3β2γ2 receptors in order to identify compounds that were selective to the

ε-containing receptor. A similar practice was also used by Wafford et al. (2009) who sought

δ-subunit-selective compounds.

The results obtained from the preliminary screening tests demonstrated that most compounds

that were active on α1β2ε receptors inhibited the receptors activity, whereas it was the op-

posite for cells expressing α3β2γ2 receptors (most active compounds enhanced the fluores-

cence quench). Why most of the compounds inhibited the fluorescence quench is not known.

It could again have something to do with the spontaneous activity of ε-containing receptors.

Next, compounds that showed an effect were further tested using additional controls. By this

means, false positive compounds could be eliminated. Several compounds showed effects

on cells that did not express the GABAA receptor subtypes, but only YFP-H148Q/I152L,

indicating that changes of the fluorescence signal were due to endogenously expressed pro-

teins rather than the GABAA receptors investigated. Endogenously expressed proteins in

HEK293 cells also likely played a role when setting up the iodide-flux assay that was tested

before the mutant YFP-based assay, but failed to give the desired outcome. The data obtained

highlighted the importance of controls.

From the 800 compounds that were screened against the two receptor subtypes, only a few

were identified as hit compounds. In order to verify their activity, the compounds were tested

using the two-electrode voltage-clamp method and Xenopus laevis oocytes expressing the

respective receptor subtypes. Due to time restrictions only few compounds could be tested.

Using oocytes expressing α3β2γ2 receptors confirmed the enhancing effect of compounds

171


(32) and (38) that were also seen using the plate reader assay. To what degree the results of

the two methods were comparable beyond that remains unclear, since the YFP-based assay

was only applied using one concentration of the test compounds. Results obtained from

oocytes expressing α1β2ε receptors could not confirm the activity of the two hit compounds

tested. It may be that the assays signal range was too low and the assay indeed was unsuitable

to screen compounds against this receptor subtype. However, currents obtained using the

two-electrode voltage-clamp technique and oocytes expressing α1β2ε receptors were very

small. It is, therefore, likely that a different test using HEK293 cells and the more sensitive

patch-clamp method could produce results that would confirm the effect of the previously

identified hit compounds.

This study as well as the work by Gilbert et al. (2009b) and Kruger et al. (2005) showed that

it is possible to set up a cell-based screening assay and screen chemical compound libraries

using cells transiently expressing mutant YFP-H148Q/I152L and certain GABAA receptor

subtypes. However, the studies also show that there is further room for improvement of the

method in the future. Gilbert et al. (2009b), for example, showed that different transient

transfection methods and transfection efficacies influence the YFP % fluorescence quench

as well as the GABA EC50 values (note that the study was performed using flow cytometry

rather than a fluorescence plate reader). The study by Kruger et al. (2005) and the work

performed here also indicated inconsistencies in conjunction with the GABA EC50 value

obtained using the fluorescence plate reader.

6.4 Future Work

This work demonstrated that the mutant YFP-based assay is suitable to screen chemical

compound libraries. As indicated above, there are still difficulties associated with the method

and no ε-selective compounds could be identified yet. However, the work presents the basis

to further search for ε-subunit-selective modulatory compounds. The improvement of the

assay’s quality and robustness could enhance the chances to identify hit compounds greatly.

One factor that might have affected the small YFP quench was the fact that ε-containing

172


receptors exhibit constitutive activity in vitro and the question arose whether a mutation

preventing the receptor from opening spontaneously could improve the assay’s outcome. The

improved assay could be used to screen further chemical compounds against ε-containing

GABAA receptors and hit compounds could be investigated in more detail.

Nevertheless, several promising compounds that showed an effect on ε-containing GABAA

receptors have been identified in this study, which could be continued by verifying the ac-

tivity of those compounds using further tests, for example the patch-clamp method using

HEK293 cells. The next logical step would be to generate structural analogues of hit com-

pounds using parallel synthesis (van Niel et al., 2005; Wafford et al., 2009). This second-

generation library could then be screened again to identify compounds with enhanced bin-

ding affinity and selectivity. Improved compounds could then be examined in more detail

using electrophysiological methods. If the activity of those compounds could be confir-

med, they would present important research tools to further study the ε-subunit-containing

GABAA receptor and help to finally elucidate the subunit’s function in vivo.

6.5 Conclusions

The work conducted showed that α3β2ε GABAA receptors exhibit similar pharmacological

characteristics as other ε-subunit-containing receptor subtypes (Davies et al., 1997; Whiting

et al., 1997; Thompson et al., 1998; Neelands et al., 1999; Maksay et al., 2003; Ranna et al.,

2006). Inclusion of the ε subunit into the α3β2 receptor complex conferred distinct pharma-

cological properties as well as unusual characteristics, such as spontaneous activity, to the

receptor subtype.

A cell-based screening assay was established using the mutant YFP-H148Q/I152L expressed

in HEK293 cells also expressing α1β2ε or α3β2γ2 receptors. It was demonstrated that

the assay format facilitated the screening for compounds that modulate the activity of these

receptor subtypes and several interesting hit compounds could be identified, of which several

have been further investigated using the two-electrode voltage-clamp technique.

173


Further work is needed to unequivocally identify ε-subunit-selective compounds. Such com-

pounds are important research tools that would contribute greatly towards the understanding

of the physiological functions of ε-containing GABAA receptor subtypes in the future. In

addition, they are prospective candidates for drug development.

174

Bibliography

Abeliovich, H. (2005). An empirical extremum principle for the Hill coefficient in ligand-protein interactions showing negative cooperativity. Biophysical Journal, 89, 76–79.

Agid, Y., Buzsaki, G., Diamond, D. M., Frackowiak, R., Giedd, J., Girault, J.-A., Grace,A., Lambert, J. J., Manji, H., Mayberg, H., Popoli, M., Prochiantz, A., Richter-Levin, G.,Somogyi, P., Spedding, M., Svenningsson, P., & Weinberger, D. (2007). How can drugdiscovery for psychiatric disorders be improved? Nature Reviews Drug Discovery, 6,189–201.

An, W. F. & Tolliday, N. J. (2009). Introduction: Cell-based assays for high-throughputscreening. In Cell-Based Assays for High-Throughput Screening, vol. 486. (HumanaPress), pp. 1–12.

Atack, J. R., Maubach, K. A., Wafford, K. A., O’Connor, D., Rodrigues, A. D., Evans,D. C., Tattersall, F. D., Chambers, M. S., MacLeod, A. M., Eng, W.-S., Ryan, C.,Hostetler, E., Sanabria, S. M., Gibson, R. E., Krause, S., Burns, H. D., Hargreaves,R. J., Agrawal, N. G. B., McKernan, R. M., Murphy, M. G., Gingrich, K., Daw-son, G. R., Musson, D. G., & Petty, K. J. (2009). In Vitro and in Vivo proper-ties of 3-tert-butyl-7-(5-methylisoxazol-3-yl)-2-(1-methyl-1h-1,2,4-triazol-5-ylmethoxy)-pyrazolo[1,5-d]-[1,2,4]triazine (MRK-016), a GABA-A receptor alpha5 subtype-selectiveinverse agonist. Journal of Pharmacology and Experimental Therapeutics, 331, 470–484.

Atack, J. R., Wafford, K. A., Tye, S. J., Cook, S. M., Sohal, B., Pike, A.,Sur, C., Melillo, D., Bristow, L., Bromidge, F., Ragan, I., Kerby, J., Street, L.,Carling, R., Castro, J. L., Whiting, P., Dawson, G. R., & McKernan, R. M.(2006). TPA023 [7-(1,1-dimethylethyl)-6-(2-ethyl-2h-1,2,4-triazol-3-ylmethoxy)-3-(2-fluorophenyl)-1,2,4-triazolo[4,3-b]pyridazine], an agonist selective for alpha2- andalpha3-containing GABA-A receptors, is a nonsedating anxiolytic in rodents and primates.Journal of Pharmacology and Experimental Therapeutics, 316, 410–422.

Bateson, A. N. (2004). The benzodiazepine site of the GABA-A receptor: an old target withnew potential? Sleep Medicine, 5, S9–S15.

Bateson, A. N., Lasham, A., & Darlison, M. G. (1991). gamma-Aminobutyric acidA receptorheterogeneity is increased by alternative splicing of a novel beta-subunit gene transcript.Journal of Neurochemistry, 56, 1437–1440.

Baumann, S. W., Baur, R., & Sigel, E. (2002). Forced subunit assembly in alpha1 beta2gamma2 GABA-A receptors. Insight into the absolute arrangement. The Journal of Bio-logical Chemistry, 277, 46020–46025.

175

BIBLIOGRAPHY

Baumgartner, W., Islas, L., & Sigworth, F. J. (1999). Two-microelectrode voltage clamp ofXenopus oocytes: Voltage errors and compensation for local current flow. BiophysicalJournal, 77, 1980–1991.

Belelli, D., Callachan, H., Hill-Venning, C., Peters, J., & Lambert, J. (1996). Interaction ofpositive allosteric modulators with human and Drosophila recombinant GABA receptorsexpressed in Xenopus laevis oocytes. British Journal of Pharmacology, 118, 563–576.

Belelli, D., Casula, A., Ling, A., & Lambert, J. J. (2002). The influence of subunit compo-sition on the interaction of neurosteroids with GABA-A receptors. Neuropharmacology,43, 651–661.

Belelli, D. & Lambert, J. J. (2005). Neurosteroids: endogenous regulators of the GABA-Areceptor. Nature Reviews Neuroscience, 6, 565–575.

Belelli, D., Lambert, J. J., Peters, J. A., Wafford, K., & Whiting, P. J. (1997). The interactionof the general anesthetic etomidate with the gamma-aminobutyric acid type A receptor isinfluenced by a single amino acid. Proceedings of the National Academy of Sciences, 94,11031–11036.

Belujon, P., Baufreton, J., Grandoso, L., Boue-Grabot, E., Batten, T., Ugedo, L., Garret, M.,& Taupignon, A. (2009). Inhibitory transmission in locus coeruleus neurons expressingGABA-A receptor epsilon subunit has a number of unique properties. Journal of Neuro-physiology, 102, 2312–2325.

Benke, D., Fakitsas, P., Roggenmoser, C., Michel, C., Rudolph, U., & Möhler, H. (2004).Analysis of the presence and abundance of GABA-A receptors containing two differenttypes of alpha subunits in murine brain using point-mutated alpha subunits. Journal ofBiological Chemistry, 279, 43654–43660.

Berglund, K., Schleich, W., Wang, H., Feng, G., Hall, W., Kuner, T., & Augustine, G.(2008). Imaging synaptic inhibition throughout the brain via genetically targeted Clo-meleon. Brain Cell Biology, 36, 101–118.

Boileau, A. J., Baur, R., Sharkey, L. M., Sigel, E., & Czajkowski, C. (2002). The relativeamount of cRNA coding for [gamma]2 subunits affects stimulation by benzodiazepines inGABA-A receptors expressed in Xenopus oocytes. Neuropharmacology, 43, 695–700.

Boileau, A. J., Li, T., Benkwitz, C., Czajkowski, C., & Pearce, R. A. (2003). Effects of[gamma]2s subunit incorporation on GABA-A receptor macroscopic kinetics. Neuro-pharmacology, 44, 1003–1012.

Bollan, K. A., Baur, R., Hales, T. G., Sigel, E., & Connolly, C. N. (2008). The promis-cuous role of the epsilon subunit in GABA-A receptor biogenesis. Molecular and CellularNeuroscience, 37, 610–621.

Bonnert, T. P., McKernan, R. M., Farrar, S., le Bourdelles, B., Heavens, R. P., Smith, D. W.,Hewson, L., Rigby, M. R., Sirinathsinghji, D. J. S., Brown, N., Wafford, K. A., & Whiting,P. J. (1999). Theta, a novel gamma-aminobutyric acid type A receptor subunit. Procee-dings of the National Academy of Sciences of the United States of America, 96, 9891–9896.

176

BIBLIOGRAPHY

Bormann, J. (2000). The ‘ABC’ of GABA receptors. Trends in Pharmacological Sciences,21, 16–19.

Bregestovski, P., Waseem, T., & Mukhtarov, M. (2009). Genetically encoded optical sensorsfor monitoring of intracellular chloride and chloride-selective channel activity. Frontiersin Molecular Neuroscience, 2, 15.

Buddhala, C., Hsu, C.-C., & Wu, J.-Y. (2009). A novel mechanism for GABA synthesis andpackaging into synaptic vesicles. Neurochemistry International, 55, 9–12.

Buffett-Jerrott, S. & Stewart, S. (2002). Cognitive and sedative effects of benzodiazepineuse. Current Pharmaceutical Design, 8, 45–58.

Celentano, J. & Wong, R. (1994). Multiphasic desensitization of the GABA-A receptor inoutside-out patches. Biophysical Journal, 66, 1039–1050.

Cervetto, C. & Taccola, G. (2008). GABA-A and strychnine-sensitive glycine receptorsmodulate N-methyl-d-aspartate-evoked acetylcholine release from rat spinal motoneurons:A possible role in neuroprotection. Neuroscience, 154, 1517–1524.

Chambers, M. S., Atack, J. R., Carling, R. W., Collinson, N., Cook, S. M., Dawson, G. R.,Ferris, P., Hobbs, S. C., O’Connor, D., Marshall, G., Rycroft, W., & MacLeod, A. M.(2004). An orally bioavailable, functionally selective inverse agonist at the benzodiaze-pine site of GABA-A alpha5 receptors with cognition enhancing properties. Journal ofMedicinal Chemistry, 47, 5829–5832.

Chen, C.-L., Yang, Y.-R., & Chiu, T.-H. (1999). Activation of rat locus coeruleus neuronGABA-A receptors by propofol and its potentiation by pentobarbital or alphaxalone. Eu-ropean Journal of Pharmacology, 386, 201–210.

Citron, M. (2010). Alzheimer’s disease: strategies for disease modification. Nature ReviewsDrug Discovery, 9, 387–398.

Colquhoun, D. (1998). Binding, gating, affinity and efficacy: the interpretation of structure-activity relationships for agonists and of the effects of mutating receptors. British Journalof Pharmacology, 125, 924–947.

Connolly, C. N., Krishek, B. J., McDonald, B. J., Smart, T. G., & Moss, S. J. (1996). As-sembly and cell surface expression of heteromeric and homomeric gamma-aminobutyricacid type A receptors. Journal of Biological Chemistry, 271, 89–96.

Crestani, F., Martin, J. R., Möhler, H., & Rudolph, U. (2000). Mechanism of action of thehypnotic zolpidem in vivo. British Journal of Pharmacology, 131, 1251–1254.

Darlison, M. G., Pahal, I., & Thode, C. (2005). Consequences of the evolution of the GABA-A receptor gene family. Cellular and Molecular Neurobiology, 25, 607–624.

Dascal, N. (1987). The use of Xenopus oocytes for the study of ion channels. CRC CriticalReviews in Biochemistry, 22, 317–387.

Davies, P. A., Hanna, M. C., Hales, T. G., & Kirkness, E. F. (1997). Insensitivity to anaes-thetic agents conferred by a class of GABA-A receptor subunit. Nature, 385, 820–823.

177

BIBLIOGRAPHY

Davies, P. A., Kirkness, E. F., & Hales, T. G. (2001). Evidence for the formation of functio-nally distinct αβγε GABA-A receptors. Journal of Physiology, 537, 101–113.

Davies, P. A., McCartney, M. R., Wang, W., Hales, T. G., & Kirkness, E. F. (2002). Alterna-tive transcripts of the GABA-A receptor epsilon subunit in human and rat. Neuropharma-cology, 43, 467–475.

D’Hulst, C., Atack, J. R., & Kooy, R. F. (2009). The complexity of the GABA-A receptorshapes unique pharmacological profiles. Drug Discovery Today, 14, 866–875.

D’Hulst, C. & Kooy, R. F. (2007). The GABA-A receptor: a novel target for treatment offragile X? Trends in Neurosciences, 30, 425–431.

Dominguez-Perrot, C., Feltz, P., & Poulter, M. O. (1996). Recombinant GABA-A receptordesensitization: the role of the gamma2 subunit and its physiological significance. TheJournal of Physiology, 497, 145–159.

Draguhn, A., Verdorn, T. A., Ewert, M., Seeburg, P. H., & Sakmann, B. (1990). Functionaland molecular distinction between recombinant rat GABA-A receptor subtypes by Zn2+.Neuron, 5, 781–788.

Ducic, I., Caruncho, H. J., Zhu, W. J., Vicini, S., & Costa, E. (1995). gamma-Aminobutyricacid gating of Cl- channels in recombinant GABA-A receptors. Journal of Pharmacologyand Experimental Therapeutics, 272, 438–445.

Duebel, J., Haverkamp, S., Schleich, W., Feng, G., Augustine, G. J., Kuner, T., & Euler, T.(2006). Two-photon imaging reveals somatodendritic chloride gradient in retinal ON-typebipolar cells expressing the biosensor Clomeleon. Neuron, 49, 81–94.

Dumont, J. (1972). Oogenesis in Xenopus laevis (Daudin). I. stages of oocyte developmentin laboratory maintained animals. Journal of Morphology, 136, 153–179.

Ebert, B., Wafford, K. A., Whiting, P. J., Krogsgaard-Larsen, P., & Kemp, J. A. (1994).Molecular pharmacology of gamma-aminobutyric acid type A receptor agonists and par-tial agonists in oocytes injected with different alpha, beta, and gamma receptor subunitcombinations. Molecular Pharmacology, 46, 957–963.

Erlitzki, R., Gong, Y., Zhang, M., & Minuk, G. (2000). Identification of gamma-aminobutyric acid receptor subunit types in human and rat liver. American Journal ofPhysiology - Gastrointestinal and Liver Physiology, 279, G733–739.

Ferri, C. P., Prince, M., Brayne, C., Brodaty, H., Fratiglioni, L., Ganguli, M., Hall, K.,Hasegawa, K., Hendrie, H., Huang, Y., Jorm, A., Mathers, C., Menezes, P. R., Rimmer,E., & Scazufca, M. (2005). Global prevalence of dementia: a Delphi consensus study. TheLancet, 366, 2112–2117.

Filatov, G. N. & White, M. M. (1995). The role of conserved leucines in the M2 domain ofthe acetylcholine receptor in channel gating. Molecular Pharmacology, 48, 379–384.

Findlay, G. S., Ueno, S., Harrison, N. L., & Harris, R. A. (2001). Allosteric modulationin spontaneously active mutant [gamma]-aminobutyric acidA receptors in frogs. Neuros-cience Letters, 293, 155–158.

178

BIBLIOGRAPHY

Fisher, J. & Macdonald, R. (1997). Functional properties of recombinant GABA-A receptorscomposed of single or multiple beta subunit subtypes. Neuropharmacology, 36, 1601–1610.

Fritschy, J.-M. & Brünig, I. (2003). Formation and plasticity of GABAergic synapses: phy-siological mechanisms and pathophysiological implications. Pharmacology & Therapeu-tics, 98, 299–323.

Galietta, L. J. V., Haggie, P. M., & Verkman, A. S. (2001a). Green fluorescent protein-basedhalide indicators with improved chloride and iodide affinities. Federation of EuropeanBiochemical Societies Letters, 499, 220–224.

Galietta, L. V. J., Jayaraman, S., & Verkman, A. S. (2001b). Cell-based assay for high-throughput quantitative screening of CFTR chloride transport agonists. American Journalof Physiology - Cell Physiology, 281, C1734–1742.

Gao, B., Fritschy, J., Benke, D., & Möhler, H. (1993). Neuron-specific expression of GABA-A-receptor subtypes: Differential association of the [alpha]1- and [alpha]3-subunits withserotonergic and GABAergic neurons. Neuroscience, 54, 881–892.

Gao, B., Hornung, J. P., & Fritschy, J. M. (1995). Identification of distinct GABA-A-receptorsubtypes in cholinergic and parvalbumin-positive neurons of the rat and marmoset medialseptum-diagonal band complex. Neuroscience, 65, 101–117.

Garret, M., Bascles, L., Boue-Grabot, E., Sartor, P., Charron, G., Bloch, B., & Margolskee,R. (1997). An mRNA encoding a putative GABA-gated chloride channel is expressed inthe human cardiac conduction system. Journal of Neurochemistry, 68, 1382–1389.

Gilbert, D., Esmaeili, A., & Lynch, J. W. (2009a). Optimizing the expression of recombi-nant αβγ GABA-A receptors in HEK293 cells for high-throughput screening. Journal ofBiomolecular Screening, 14, 86–91.

Gilbert, D. F., Islam, R., Lynagh, T. P., Webb, T. I., & Lynch, J. (2009b). High through-put techniques for discovering new glycine receptor modulators and their binding sites.Frontiers in Molecular Neuroscience, 2, 1–10.

Gill, S., Gill, R., Lee, S., Hesketh, J., Fedida, D., Rezazadeh, S., Stankovich, L., & Liang,D. (2003). Flux assays in high throughput screening of ion channels in drug discovery.ASSAY and Drug Development Technologies, 1, 709–717.

Gillet, V. J. (2008). New directions in library design and analysis. Current Opinion inChemical Biology, 12, 372–378.

Gingrich, K. J., Roberts, W. A., & Kass, R. S. (1995). Dependence of the GABA-A receptorgating kinetics on the alpha-subunit isoform: implications for structure-function relationsand synaptic transmission. The Journal of Physiology, 489, 529–543.

Gonzalez, J. & Maher, M. (2002). Cellular fluorescent indicators and voltage/ion probereader (VIPR) tools for ion channel and receptor drug discovery. Receptors and Channels,8, 283–295.

179

BIBLIOGRAPHY

Haas, K. & Macdonald, R. (1999). GABA-A receptor subunit gamma2 and delta subtypesconfer unique kinetic properties on recombinant GABA-A receptor currents in mouse fi-broblasts. The Journal of Physiology, 514, 27–45.

Hadingham, K. L., Wingrove, P., Bourdelles Le, B., Palmer, K. J., Ragan, C. I., & Whiting,P. J. (1993). Cloning of cDNA sequences encoding human alpha2 and alpha3 gamma-aminobutyric acidA receptor subunits and characterization of the benzodiazepine pharma-cology of recombinant alpha1-, alpha2-, alpha3- , and alpha5-containing human gamma-aminobutyric acidA receptors. Molecular Pharmacology, 43, 970–975.

Hamill, O., Marty, A., Neher, E., Sakmann, B., & Sigworth, F. (1981). Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membranepatches. Pfluegers Archiv: European Journal of Physiology, 391, 85–100.

Hann, M. M. & Oprea, T. I. (2004). Pursuing the leadlikeness concept in pharmaceuticalresearch. Current Opinion in Chemical Biology, 8, 255–263.

Harris, B. D., Wong, G., Moody, E. J., & Skolnick, P. (1995). Different subunit requirementsfor volatile and nonvolatile anesthetics at gamma-aminobutyric acid type A receptors. Mo-lecular Pharmacology, 47, 363–367.

Harris, R., Mihic, S., Brozowski, S., Hadingham, K., & Whiting, P. (1997). Ethanol, flunitra-zepam, and pentobarbital modulation of GABA-A receptors expressed in mammalian cellsand Xenopus oocytes. Alcoholism: Clinical and Experimental Research, 21, 444–451.

Harvey, R. J., Kim, H.-C., & Darlison, M. G. (1993). Molecular cloning reveals the existenceof a fourth [gamma] subunit of the vertebrate brain GABA-A receptor. Federation ofEuropean Biochemical Societies Letters, 331, 211–216.

Haverkamp, S., Wassle, H., Duebel, J., Kuner, T., Augustine, G. J., Feng, G., & Euler, T.(2005). The primordial, blue-cone color system of the mouse retina. The Journal ofNeuroscience, 25, 5438–5445.

Hedblom, E. & Kirkness, E. F. (1997). A novel class of GABA-A receptor subunit in tissuesof the reproductive system. Journal of Biological Chemistry, 272, 15346–15350.

Hevers, W. & Lüddens, H. (1998). The diversity of GABA-A receptors. Pharmacological andelectrophysiological properties of GABA-A channel subtypes. Molecular Neurobiology,18, 35–86.

Hill-Venning, C., Belelli, D., Peters, J., & Lambert, J. (1997). Subunit-dependent interactionof the general anaesthetic etomidate with the gamma-aminobutyric acid type A receptor.British Journal of Pharmacology, 120, 749–756.

Hogg, R. C., Bandelier, F., Benoit, A., Dosch, R., & Bertrand, D. (2008). An automatedsystem for intracellular and intranuclear injection. Journal of Neuroscience Methods, 169,65–75.

Hosie, A. M., Clarke, L., da Silva, H., & Smart, T. G. (2009). Conserved site for neurosteroidmodulation of GABA-A receptors. Neuropharmacology, 56, 149–154.

Hosie, A. M., Wilkins, M. E., & Smart, T. G. (2007). Neurosteroid binding sites on GABA-Areceptors. Pharmacology & Therapeutics, 116, 7–19.

180

BIBLIOGRAPHY

Inglese, J., Auld, D. S., Jadhav, A., Johnson, R. L., Simeonov, A., Yasgar, A., Zheng, W., &Austin, C. P. (2006). Quantitative high-throughput screening: A titration-based approachthat efficiently identifies biological activities in large chemical libraries. Proceedings ofthe National Academy of Sciences, 103, 11473–11478.

Inglese, J., Johnson, R. L., Simeonov, A., Xia, M., Zheng, W., Austin, C. P., & Auld, D. S.(2007a). High-throughput screening assays for the identification of chemical probes. Na-ture Chemical Biology, 3, 466–479.

Inglese, J., Shamu, C. E., & Guy, R. K. (2007b). Reporting data from high-throughputscreening of small-molecule libraries. Nature Chemical Biology, 3, 438–441.

Irnaten, M., Walwyn, W., Wang, J., Venkatesan, P., Evans, C., Chang, K., Andresen, M.,Hales, T., & Mendelowitz, D. (2002). Pentobarbital enhances GABAergic neurotransmis-sion to cardiac parasympathetic neurons, which is prevented by expression of GABA-Aepsilon subunit. Anesthesiology, 97, 717–724.

Janzen, W. P. & Popa-Burke, I. G. (2009). Advances in improving the quality and flexibilityof compound management. Journal of Biomolecular Screening, 14, 444–451.

Jayaraman, S., Haggie, P., Wachter, R. M., Remington, S. J., & Verkman, A. S. (2000).Mechanism and cellular applications of a green fluorescent protein-based halide sensor.The Journal of Biological Chemistry, 275, 6047–6050.

Jechlinger, M., Pelz, R., Tretter, V., Klausberger, T., & Sieghart, W. (1998). Subunit com-position and quantitative importance of hetero-oligomeric receptors: GABA-A receptorscontaining alpha6 subunits. The Journal of Neuroscience, 18, 2449–2457.

Johnstone, T. B. C., Hogenkamp, D. J., Coyne, L., Su, J., Halliwell, R. F., Tran, M. B., Yo-shimura, R. F., Li, W.-Y., Wang, J., & Gee, K. W. (2004). Modifying quinolone antibioticsyields new anxiolytics. Nature Medicine, 10, 31–32.

Jones, B. L. & Henderson, L. P. (2007). Trafficking and potential assembly patterns ofepsilon-containing GABA-A receptors. Journal of Neurochemistry, 103, 1258–1271.

Jones, B. L., Whiting, P. J., & Henderson, L. P. (2006). Mechanisms of anabolic androge-nic steroid inhibition of mammalian epsilon-subunit-containing GABA-A receptors. TheJournal of Physiology, 573, 571–593.

Jurd, R., Arras, M., Lambert, S., Drexler, B., Siegwart, R., Crestani, F., Zaugg, M., Vogt,K. E., Ledermann, B., Antkowiak, B., & Rudolph, U. (2003). General anesthetic actionsin vivo strongly attenuated by a point mutation in the GABA-A receptor beta3 subunit. TheJournal of the Federation of American Societies for Experimental Biology, 17, 250–252.

Kar, S., Slowikowski, S., Westaway, D., & Mount, H. (2004). Interactions between beta-amyloid and central cholinergic neurons: implications for Alzheimer’s disease. Journal ofPsychiatry and Neuroscience, 29, 427–441.

Kash, T. L., Trudell, J. R., & Harrison, N. L. (2004). Structural elements involved in acti-vation of the gamma-aminobutyric acid type A (GABA-A) receptor. Biochemical SocietyTransactions, 32, 540–546.

181

BIBLIOGRAPHY

Kasparov, S., Davies, K., Patel, U., Boscan, P., Garret, M., & Paton, J. (2001). GABA-Areceptor epsilon-subunit may confer benzodiazepine insensitivity to the caudal aspect ofthe nucleus tractus solitarii of the rat. The Journal of Physiology, 536, 785–96.

Kaufman, D. L., Houser, C. R., & Tobin, A. J. (1991). Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronaldistributions and cofactor interactions. Journal of Neurochemistry, 56, 720–723.

Kim, H.-J., Yun, H.-M., Kim, T., Nam, G., Roh, E. J., Kostenis, E., Choo, H.-Y. P., Pae,A. N., & Rhim, H. (2008). Functional human 5-HT6 receptor assay for high throughputscreening of chemical ligands and binding proteins. Combinatorial Chemistry & HighThroughput Screening, 11, 316–324.

Knoflach, F., Rhyner, T., Villa, M., Kellenberger, S., Drescher, U., Malherbe, P., Sigel, E., &Möhler, H. (1991). The [gamma]3-subunit of the GABA-A-receptor confers sensitivity tobenzodiazepine receptor ligands. Federation of European Biochemical Societies Letters,293, 191–194.

Korpi, E. R., Gründer, G., & Lüddens, H. (2002). Drug interactions at GABA-A receptors.Progress in Neurobiology, 67, 113–159.

Korpi, E. R., Kuner, T., Kristo, P., Kohler, M., Herb, A., Lüddens, H., & Seeburg, P. H.(1994). Small N-terminal deletion by splicing in cerebellar alpha6 subunit abolishesGABA-A receptor function. Journal of Neurochemistry, 63, 1167–1170.

Krishek, B. J., Moss, S. J., & Smart, T. G. (1998). Interaction of H+ and Zn2+ on recombinantand native rat neuronal GABA-A receptors. The Journal of Physiology, 507, 639–652.

Kruger, W., Gilbert, D., Hawthorne, R., Hryciw, D. H., Frings, S., Poronnik, P., & Lynch,J. W. (2005). A yellow fluorescent protein-based assay for high-throughput screening ofglycine and GABA-A receptor chloride channels. Neuroscience Letters, 380, 340–345.

Kuner, T. & Augustine, G. J. (2000). A genetically encoded ratiometric indicator for chlo-ride: Capturing chloride transients in cultured hippocampal neurons. Neuron, 27, 447–459.

Labarca, C., Nowak, M. W., Zhang, H., Tang, L., Deshpande, P., & Lester, H. A. (1995).Channel gating governed symmetrically by conserved leucine residues in the M2 domainof nicotinic receptors. Nature, 376, 514–516.

Lamigeon, C., Bellier, J. P., Sacchettoni, S., Rujano, M., & Jacquemont, B. (2001). Enhancedneuronal protection from oxidative stress by co-culture with glutamic acid decarboxylase-expressing astrocytes. Journal of Neurochemistry, 77, 598–606.

Lavoie, A., Tingey, J., Harrison, N., Pritchett, D., & Twyman, R. (1997). Activation anddeactivation rates of recombinant GABA-A receptor channels are dependent on alpha-subunit isoform. Biophysical Journal, 73, 2518–2526.

Lipinski, C. A., Lombardo, F., Dominy, B. W., & Feeney, P. J. (2001). Experimental andcomputational approaches to estimate solubility and permeability in drug discovery anddevelopment settings. Advanced Drug Delivery Reviews, 46, 3–26.

182

BIBLIOGRAPHY

Löw, K., Crestani, F., Keist, R., Benke, D., Brunig, I., Benson, J. A., Fritschy, J.-M., Rulicke,T., Bluethmann, H., Möhler, H., & Rudolph, U. (2000). Molecular and neuronal substratefor the selective attenuation of anxiety. Science, 290, 131–134.

Lüscher, B. & Keller, C. A. (2004). Regulation of GABA-A receptor trafficking, channelactivity, and functional plasticity of inhibitory synapses. Pharmacology & Therapeutics,102, 195–221.

Lynch, J. (2005). High-throughput screening of neuronal Cl- channels: Why and how?Current Neuropharmacology, 3, 207–216.

Lynch, J. W. (2009). Native glycine receptor subtypes and their physiological roles. Neuro-pharmacology, 56, 303–309.

Macdonald, R. L., Kang, J.-Q., & Gallagher, M. J. (2010). Mutations in GABA-A receptorsubunits associated with genetic epilepsies. The Journal of Physiology, 588, 1861–1869.

Maksay, G., Thompson, S. A., & Wafford, K. A. (2003). The pharmacology of spontaneouslyopen alpha1 beta3 epsilon GABA-A receptor-ionophores. Neuropharmacology, 44, 994–1002.

Maniatis, T., Fritsch, E., & Sambrook, J. (1982). Molecular Cloning: a laboratory manual.(New York: Cold Spring Harbor Laboratory Press.).

Marandi, N., Konnerth, A., & Garaschuk, O. (2002). Two-photon chloride imaging in neu-rons of brain slices. Pflügers Archiv European Journal of Physiology, 445, 357–365.

Markova, O., Mukhtarov, M., Real, E., Jacob, Y., & Bregestovski, P. (2008). Geneticallyencoded chloride indicator with improved sensitivity. Journal of Neuroscience Methods,170, 67–76.

Martin, D. L. & Rimvall, K. (1993). Regulation of gamma-aminobutyric acid synthesis inthe brain. Journal of Neurochemistry, 60, 395–407.

McGehee, D. S. & Role, L. W. (1996). Presynaptic ionotropic receptors. Current Opinion inNeurobiology, 6, 342–349.

McKernan, R. M. & Whiting, P. J. (1996). Which GABA-A-receptor subtypes really occurin the brain? Trends in Neurosciences, 19, 139–143.

Miko, A., Werby, E., Sun, H., Healey, J., & Zhang, L. (2004). A TM2 residue in thebeta1 subunit determines spontaneous opening of homomeric and heteromeric gamma-aminobutyric acid-gated ion channels. Journal of Biological Chemistry, 279, 22833–22840.

Minier, F. & Sigel, E. (2004). Techniques: Use of concatenated subunits for the study ofligand-gated ion channels. Trends in Pharmacological Sciences, 25, 499–503.

Mitchell, E. A., Herd, M. B., Gunn, B. G., Lambert, J. J., & Belelli, D. (2008). Neurosteroidmodulation of GABA-A receptors: Molecular determinants and significance in health anddisease. Neurochemistry International, 52, 588–595.

Möhler, H. (2006). GABA-A receptor diversity and pharmacology. Cell and Tissue Re-search, 326, 505–516.

183

BIBLIOGRAPHY

Möhler, H. (2007). Molecular regulation of cognitive functions and developmental plasticity:impact of GABA-A receptors. Journal of Neurochemistry, 102, 1–12.

Möhler, H., Fritschy, J. M., & Rudolph, U. (2002). A new benzodiazepine pharmacology.The Journal of Pharmacology and Experimental Therapeutics, 300, 2–8.

Molokanova, E. & Savchenko, A. (2008). Bright future of optical assays for ion channeldrug discovery. Drug Discovery Today, 13, 14–22.

Moragues, N., Ciofi, P., Lafon, P., Odessa, M. F., Tramu, G., & Garret, M. (2000). cDNAcloning and expression of a gamma-aminobutyric acid A receptor epsilon-subunit in ratbrain. European Journal of Neuroscience, 12, 4318–4330.

Moragues, N., Ciofi, P., Tramu, G., & Garret, M. (2002). Localisation of GABA-A receptor[epsilon]-subunit in cholinergic and aminergic neurones and evidence for co-distributionwith the [theta]-subunit in rat brain. Neuroscience, 111, 657–669.

Mortensen, M., Wafford, K. A., Wingrove, P., & Ebert, B. (2003). Pharmacology of GABA-A receptors exhibiting different levels of spontaneous activity. European Journal of Phar-macology, 476, 17–24.

Moult, P. R. (2009). Neuronal glutamate and GABA-A receptor function in health and di-sease. Biochemical Society Transactions, 37, 1317–1322.

Navarro, J. F., Burón, E., & Martín-López, M. (2002). Anxiogenic-like activity of L-655,708,a selective ligand for the benzodiazepine site of GABA-A receptors which contain thealpha-5 subunit, in the elevated plus-maze test. Progress in Neuro-Psychopharmacologyand Biological Psychiatry, 26, 1389–1392.

Neelands, T. R., Fisher, J. L., Bianchi, M., & Macdonald, R. L. (1999). Spontaneous andgamma-aminobutyric acid (GABA)-activated GABA-A receptor channels formed by ep-silon subunit-containing isoforms. Molecular Pharmacology, 55, 168–178.

Olsen, R. W., Chang, C.-S. S., Li, G., Hanchar, H. J., & Wallner, M. (2004). Fishing forallosteric sites on GABA-A receptors. Biochemical Pharmacology, 68, 1675–1684.

Olsen, R. W. & Sieghart, W. (2008). International union of pharmacology. LXX. subtypes ofgamma-aminobutyric acidA receptors: Classification on the basis of subunit composition,pharmacology, and function. update. Pharmacological Reviews, 60, 243–260.

Olsen, R. W. & Sieghart, W. (2009). GABA-A receptors: Subtypes provide diversity offunction and pharmacology. Neuropharmacology, 56, 141–148.

Pape, J. R., Bertrand, S. S., Lafon, P., Odessa, M. F., Chaigniau, M., Stiles, J. K., & Garret,M. (2009). Expression of GABA-A receptor [alpha]3-, [theta]-, and [epsilon]-subunitmRNAs during rat CNS development and immunolocalization of the [epsilon] subunit indeveloping postnatal spinal cord. Neuroscience, 160, 85–96.

Pirker, S., Schwarzer, C., Wieselthaler, A., Sieghart, W., & Sperk, G. (2000). GABA-Areceptors: immunocytochemical distribution of 13 subunits in the adult rat brain. Neuros-cience, 101, 815–850.

184

BIBLIOGRAPHY

Pistis, M., Belelli, D., Peters, J. A., & Lambert, J. J. (1997). The interaction of generalanaesthetics with recombinant GABA-A and glycine receptors expressed in Xenopus lae-vis oocytes: a comparative study. British Journal of Pharmacology, 122, 1707–1719.

Piston, D. W. & Kremers, G.-J. (2007). Fluorescent protein FRET: the good, the bad and theugly. Trends in Biochemical Sciences, 32, 407–414.

Pless, S. & Lynch, J. (2008). Illuminating the structure and function of Cys-loop receptors.Clinical and Experimental Pharmacology and Physiology, 35, 1137–1142.

Pritchett, D. B., Lüddens, H., & Seeburg, P. H. (1989a). Type I and type II GABA-A-benzodiazepine receptors produced in transfected cells. Science, 245, 1389–1392.

Pritchett, D. B. & Seeburg, P. H. (1990). Gamma-aminobutyric acidA receptor alpha 5-subunit creates novel type II benzodiazepine receptor pharmacology. Journal of Neuro-chemistry, 54, 1802–1804.

Pritchett, D. B., Sontheimer, H., Shivers, B. D., Ymer, S., Kettenmann, H., Schofield, P. R.,& Seeburg, P. H. (1989b). Importance of a novel GABA-A receptor subunit for benzodia-zepine pharmacology. Nature, 338, 582–585.

Ranna, M., Sinkkonen, S., Moykkynen, T., Uusi-Oukari, M., & Korpi, E. (2006). Impact ofepsilon and theta subunits on pharmacological properties of alpha3beta1 GABA-A recep-tors expressed in Xenopus oocytes. BioMed Central Pharmacology, 6, 1.

Reynolds, D. S., Rosahl, T. W., Cirone, J., O’Meara, G. F., Haythornthwaite, A., Newman,R. J., Myers, J., Sur, C., Howell, O., Rutter, A. R., Atack, J., Macaulay, A. J., Hadingham,K. L., Hutson, P. H., Belelli, D., Lambert, J. J., Dawson, G. R., McKernan, R., Whiting,P. J., & Wafford, K. A. (2003). Sedation and anesthesia mediated by distinct GABA-Areceptor isoforms. The Journal of Neuroscience, 23, 8608–8617.

Rohrbough, J. & Spitzer, N. (1996). Regulation of intracellular Cl- levels by Na+-dependent Cl- cotransport distinguishes depolarizing from hyperpolarizing GABA-Areceptor-mediated responses in spinal neurons. The Journal of Neuroscience, 16, 82–91.

Rudolph, U., Crestani, F., Benke, D., Brunig, I., Benson, J. A., Fritschy, J.-M., Martin,J. R., Bluethmann, H., & Möhler, H. (1999). Benzodiazepine actions mediated by specific[gamma]-aminobutyric acidA receptor subtypes. Nature, 401, 796–800.

Rudolph, U., Crestani, F., & Möhler, H. (2001). GABA-A receptor subtypes: dissecting theirpharmacological functions. Trends in Pharmacological Sciences, 22, 188–194.

Rudolph, U. & Möhler, H. (2004). Analysis of GABA-A receptor function and dissectionof the pharmacology of benzodiazepines and general anesthetics through mouse genetics.Annual Review of Pharmacology and Toxicology, 44, 475–498.

Rudolph, U. & Möhler, H. (2006). GABA-based therapeutic approaches: GABA-A receptorsubtype functions. Current Opinion in Pharmacology, 6, 18–23.

Sandell, E. B. & Kolthoff, I. M. (1937). Micro determination of iodine by a catalytic method.Microchimica Acta, 1, 9–25.

185

BIBLIOGRAPHY

Sanna, E., Mascia, M. P., Klein, R. L., Whiting, P. J., Biggio, G., & Harris, R. A. (1995).Actions of the general anesthetic propofol on recombinant human GABA-A receptors:influence of receptor subunits. The Journal of Pharmacology and Experimental Therapeu-tics, 274, 353–360.

Sanna, E., Murgia, A., Casula, A., & Biggio, G. (1997). Differential subunit dependence ofthe actions of the general anesthetics alphaxalone and etomidate at gamma-aminobutyricacid type A receptors expressed in Xenopus laevis oocytes. Molecular Pharmacology, 51,484–490.

Saxena, N. C. & Macdonald, R. L. (1994). Assembly of GABA-A receptor subunits: role ofthe delta subunit. The Journal of Neuroscience, 14, 7077–7086.

Scheller, M. & Forman, S. A. (2002). Coupled and uncoupled gating and desensitizationeffects by pore domain mutations in GABA-A receptors. The Journal of Neuroscience,22, 8411–8421.

Schofield, P. R., Darlison, M. G., Fujita, N., Burt, D. R., Stephenson, F. A., Rodriguez, H.,Rhee, L. M., Ramachandran, J., Reale, V., Glencorse, T. A., Seeburg, P., & Barnard, E. A.(1987). Sequence and functional expression of the GABA-A receptor shows a ligand-gatedreceptor super-family. Nature, 328, 221–227.

Seltzer, B. (2010). Galantamine-ER for the treatment of mild-to-moderate Alzheimer’s di-sease. Clinical Interventions in Aging, 5, 1–6.

Sergeeva, O. A., Andreeva, N., Garret, M., Scherer, A., & Haas, H. L. (2005). Pharma-cological properties of GABA-A receptors in rat hypothalamic neurons expressing theepsilon-subunit. The Journal of Neuroscience, 25, 88–95.

Sieghart, W. (2000). Unraveling the function of GABA-A receptor subtypes. Trends inPharmacological Sciences, 21, 411–413.

Sieghart, W. & Sperk, G. (2002). Subunit composition, distribution and function of GABA-Areceptor subtypes. Current Topics in Medicinal Chemistry, 2, 795–816.

Sigel, E. (1990). Use of Xenopus oocytes for the functional expression of plasma membraneproteins. The Journal of Membrane Biology, 117, 201–221.

Sigel, E. & Baur, R. (2000). Electrophysiological evidence for the coexistence of alpha1 andalpha6 subunits in a single functional GABA-A receptor. Journal of Neurochemistry, 74,2590–2596.

Sigel, E. & Buhr, A. (1997). The benzodiazepine binding site of GABA-A receptors. Trendsin Pharmacological Sciences, 18, 425–429.

Sigel, E. & Minier, F. (2005). The Xenopus oocyte: System for the study of functionalexpression and modulation of proteins. Molecular Nutrition and Food Research, 49, 228–234.

Simon, J., Wakimoto, H., Fujita, N., Lalande, M., & Barnard, E. A. (2004). Analysis of theset of GABA-A receptor genes in the human genome. Journal of Biological Chemistry,279, 41422–41435.

186

BIBLIOGRAPHY

Sinkkonen, S. T., Hanna, M. C., Kirkness, E. F., & Korpi, E. R. (2000). GABA-A receptorepsilon and theta subunits display unusual structural variation between species and areenriched in the rat locus ceruleus. The Journal of Neuroscience, 20, 3588–3595.

Smith, A. & Simpson, P. (2003). Methodological approaches for the study of GABA-Areceptor pharmacology and functional responses. Analytical and Bioanalytical Chemistry,377, 843–851.

Smith, A. J., Alder, L., Silk, J., Adkins, C., Fletcher, A. E., Scales, T., Kerby, J., Marshall,G., Wafford, K. A., McKernan, R. M., & Atack, J. R. (2001). Effect of alpha subuniton allosteric modulation of ion channel function in stably expressed human recombinantgamma-aminobutyric acidA receptors determined using 36Cl ion flux. Molecular Pharma-cology, 59, 1108–1118.

Smith, G. B. & Olsen, R. W. (1995). Functional domains of GABA-A receptors. Trends inPharmacological Sciences, 16, 162–168.

Snodgrass, S. (1978). Use of 3H-muscimol for GABA receptor studies. Nature, 273, 392–394.

Stephenson, F. A., Duggan, M. J., & Pollard, S. (1990). The gamma 2 subunit is an integralcomponent of the gamma-aminobutyric acidA receptor but the alpha 1 polypeptide is theprincipal site of the agonist benzodiazepine photoaffinity labeling reaction. The Journalof Biological Chemistry, 265, 21160–21165.

Sternfeld, F., Carling, R. W., Jelley, R. A., Ladduwahetty, T., Merchant, K. J., Moore, K. W.,Reeve, A. J., Street, L. J., O’Connor, D., Sohal, B., Atack, J. R., Cook, S., Seabrook, G.,Wafford, K., Tattersall, F. D., Collinson, N., Dawson, G. R., Castro, J. L., & MacLeod,A. M. (2004). Selective, orally active gamma-aminobutyric acidA alpha5 receptor inverseagonists as cognition enhancers. Journal of Medicinal Chemistry, 47, 2176–2179.

Sui, Y. & Wu, Z. (2007). Alternative statistical parameter for high-throughput screeningassay quality assessment. Journal of Biomolecular Screening, 12, 229–234.

Tang, W. & Wildey, M. J. (2004). Development of a colorimetric method for functionalchloride channel assay. Journal of Biomolecular Screening, 9, 607–613.

Tang, W. & Wildey, M. J. (2006). Methods for measuring chloride channel conducti-vity. World Intellectual Property Organization: International Publication Number WO2006/009986 A1, p. 26 January.

Terstappen, G. C. (2005). Ion channel screening technologies today. Drug Discovery Today:Technologies, 2, 133–140.

Thomas, P. & Smart, T. G. (2005). HEK293 cell line: A vehicle for the expression ofrecombinant proteins. Journal of Pharmacological and Toxicological Methods, 51, 187–200.

Thompson, S., Bonnert, T., Whiting, P., & Wafford, K. (1998). Functional characteristicsof recombinant human GABA-A receptors containing the epsilon-subunit. ToxicologyLetters, 100-101, 233–238.

187

BIBLIOGRAPHY

Thompson, S., Whiting, P., & Wafford, K. (1996). Barbiturate interactions at the humanGABA-A receptor: dependence on receptor subunit combination. British Journal of Phar-macology, 117, 521–527.

Thompson, S. A., Bonnert, T. P., Cagetti, E., Whiting, P. J., & Wafford, K. A. (2002). Overex-pression of the GABA-A receptor epsilon subunit results in insensitivity to anaesthetics.Neuropharmacology, 43, 662–668.

Tsien, R. Y. (1998). The green fluorescence protein. Annual Review of Biochemistry, 67,509–544.

Ueno, S., Zorumski, C., Bracamontes, J., & Steinbach, J. H. (1996). Endogenous subunitscan cause ambiguities in the pharmacology of exogenous gamma-aminobutyric acidA re-ceptors expressed in human embryonic kidney 293 cells. Molecular Pharmacology, 50,931–938.

Uusi-Oukari, M., Heikkilä, J., Sinkkonen, S., Mäkelä, R., Hauer, B., Homanics, G., Sieghart,W., Wisden, W., & Korpi, E. (2000). Long-range interactions in neuronal gene expres-sion: Evidence from gene targeting in the GABA-A receptor [beta]2-[alpha]6-[alpha]1-[gamma]2 subunit gene cluster. Molecular and Cellular Neuroscience, 16, 34–41.

van Niel, M. B., Wilson, K., Adkins, C. H., Atack, J. R., Castro, J. L., Clarke, D. E., Fletcher,S., Gerhard, U., Mackey, M. M., Malpas, S., Maubach, K., Newman, R., O’Connor, D.,Pillai, G. V., Simpson, P. B., Thomas, S. R., & MacLeod, A. M. (2005). A new pyridazineseries of GABA-A alpha5 ligands. Journal of Medicinal Chemistry, 48, 6004–6011.

Verdoorn, T. A. (1994). Formation of heteromeric gamma-aminobutyric acid type A recep-tors containing two different alpha subunits. Molecular Pharmacology, 45, 475–480.

Verkman, A. S. & Galietta, L. J. V. (2009). Chloride channels as drug targets. NatureReviews Drug Discovery, 8, 153–171.

Waagepetersen, H. S., Sonnewald, U., & Schousboe, A. (1999). The GABA paradox: mul-tiple roles as metabolite, neurotransmitter, and neurodifferentiative agent. Journal of Neu-rochemistry, 73, 1335–1342.

Wachter, R. M., Elsliger, M.-A., Kallio, K., Hanson, G. T., & Remington, S. J. (1998). Struc-tural basis of spectral shifts in the yellow-emission variants of green fluorescent protein.Structure, 6, 1267–1277.

Wachter, R. M. & Remington, S. J. (1999). Sensitivity of the yellow variant of green fluo-rescent protein to halides and nitrate. Current Biology, 9, R628–R629.

Wafford, K. A., Bain, C. J., Whiting, P. J., & Kemp, J. A. (1993). Functional com-parison of the role of gamma subunits in recombinant human gamma-aminobutyricacidA/benzodiazepine receptors. Molecular Pharmacology, 44, 437–442.

Wafford, K. A., van Niel, M. B., Ma, Q. P., Horridge, E., Herd, M. B., Peden, D. R., Belelli,D., & Lambert, J. J. (2009). Novel compounds selectively enhance [delta] subunit contai-ning GABA-A receptors and increase tonic currents in thalamus. Neuropharmacology, 56,182–189.

188

BIBLIOGRAPHY

Wagner, C., Friedrich, B., Setiawan, I., Lang, F., & S., B. (2000). The use of Xenopuslaevis oocytes for the functional characterization of heterologously expressed membraneproteins. Cellular Physiology and Biochemistry, 10, 1–12.

Wagner, D. A., Goldschen-Ohm, M. P., Hales, T. G., & Jones, M. V. (2005). Kinetics andspontaneous open probability conferred by the epsilon subunit of the GABA-A receptor.The Journal of Neuroscience, 25, 10462–10468.

Watkins, P. B., Zimmerman, H. J., Knapp, M. J., Gracon, S. I., & Lewis, K. W. (1994).Hepatotoxic effects of tacrine administration in patients with Alzheimer’s disease. JAMA:The Journal of the American Medical Association, 271, 992–998.

Whiting, P., Bonnert, T., McKernan, R., Farrar, S., Bourdelles, B. l., Heavens, R., Smith, D.,Hewson, L., Rigby, M., Sirinathsinghji, J., Thompson, S., & Wafford, K. (1999). Mole-cular and functional diversity of the expanding GABA-A receptor gene family. Annals ofthe New York Academy of Sciences, 868, 645–653.

Whiting, P., McKernan, R. M., & Iversen, L. L. (1990). Another mechanism for creatingdiversity in gamma-aminobutyrate type A receptors: RNA splicing directs expression oftwo forms of gamma 2 phosphorylation site. Proceedings of the National Academy ofSciences of the United States of America, 87, 9966–9970.

Whiting, P. J. (1999). The GABA-A receptor gene family: new targets for therapeutic inter-vention. Neurochemistry International, 34, 387–390.

Whiting, P. J. (2006). GABA-A receptors: a viable target for novel anxiolytics? CurrentOpinion in Pharmacology, 6, 24–29.

Whiting, P. J., McAllister, G., Vassilatis, D., Bonnert, T. P., Heavens, R. P., Smith, D. W.,Hewson, L., O’Donnell, R., Rigby, M. R., Sirinathsinghji, D. J. S., Marshall, G., Thomp-son, S. A., & Wafford, K. A. (1997). Neuronally restricted RNA splicing regulates theexpression of a novel GABA-A receptor subunit conferring atypical functional properties.The Journal of Neuroscience, 17, 5027–5037.

Wilke, K., Gaul, R., Klauck, S. M., & Poustka, A. (1997). A gene in human chromosomeband Xq28 (GABRE) defines a putative new subunit class of the GABA-A neurotransmit-ter receptor. Genomics, 45, 1–10.

Willumsen, N. J., Bech, M., Olesen, S.-P., Jensen, B. S., Korsgaard, M. P. G., & Christo-phersen, P. (2003). High throughput electrophysiology: New perspectives for ion channeldrug discovery. Receptors and Channels, 9, 3–12.

Wisden, W., Herb, A., Wieland, H., Keinanen, K., Luddens, H., & Seeburg, P. H. (1991).Cloning, pharmacological characteristics and expression pattern of the rat GABA-A re-ceptor [alpha]4 subunit. Federation of European Biochemical Societies Letters, 289, 227–230.

Witchel, H. J., Milnes, J. T., Mitcheson, J. S., & Hancox, J. C. (2002). Troubleshootingproblems with in vitro screening of drugs for QT interval prolongation using HERG K+

channels expressed in mammalian cell lines and Xenopus oocytes. Journal of Pharmaco-logical and Toxicological Methods, 48, 65–80.

189

BIBLIOGRAPHY

Wohlfarth, K. M., Bianchi, M. T., & Macdonald, R. L. (2002). Enhanced neurosteroidpotentiation of ternary GABA-A receptors containing the delta subunit. The Journal ofNeuroscience, 22, 1541–1549.

Wong, G., Sei, Y., & Skolnick, P. (1992). Stable expression of type I gamma-aminobutyricacidA/benzodiazepine receptors in a transfected cell line. Molecular Pharmacology, 42,996–1003.

Wood, C., Williams, C., & Waldron, G. J. (2004). Patch clamping by numbers. DrugDiscovery Today, 9, 434–441.

Xu, J., Wang, X., Ensign, B., Li, M., Wu, L., Guia, A., & Xu, J. (2001). Ion-channel assaytechnologies: quo vadis? Drug Discovery Today, 6, 1278–1287.

Yuede, C. M., Donga, H., & Csernansky, J. G. (2007). Anti-dementia drugs andhippocampal-dependent memory in rodents. Behavioural Pharmacology, 18, 347–363.

Zhang, J., Sato, M., & Tohyama, M. (1991). Region-specific expression of the mRNAsencoding beta subunits (beta 1, beta 2, and beta 3) of GABA-A receptor in the rat brain.The Journal of Comparative Neurology, 303, 637–57.

Zhang, J.-H., Chung, T. D. Y., & Oldenburg, K. R. (1999). A simple statistical parame-ter for use in evaluation and validation of high throughput screening assays. Journal ofBiomolecular Screening, 4, 67–73.

Zheng, W., Spencer, R. H., & Kiss, L. (2004). High throughput assay technologies for ionchannel drug discovery. ASSAY and Drug Development Technologies, 2, 543–552.

190

Appendices

191

Appendix A

Vector maps

Figure A.1: Vector map pcDNA3.1 (Invitrogen)

192

Chapter A. Vector maps

Figure A.2: Vector map pcDNA3 (Invitrogen)

Figure A.3: Vector map pCMV (from Professor Erwin Sigel)

193

Chapter A. Vector maps

Figure A.4: Vector map pBS (Stratagene)

Figure A.5: Vector map pRK5-Clomeleon (from Dr Thomas Kuner)

194

Appendix B

Additional Graphs

GABA

GABA + 0.1% DMSO

0.1% DMSO

GABA + 0.2% DMSO

0.2% DMSO

0

20

40

60

80

100

α3β2εα3β2γ2

α3β2

% G

AB

A c

on

tro

l res

po

nse

Figure B.1: Effect of DMSO on oocytes expressing α3β2, α3β2ε and α3β2γ2 GABAAreceptor subtypes.Columns represent mean responses of different concentrations of DMSO alone or with GABA (EC20concentrations) on the receptor subtypes. α3β2: n = 4 for GABA-control response, n = 3 for GABA+ 0.1 % DMSO, n = 3 for 0.1 % DMSO, n = 3 for GABA + 0.2 % DMSO, n = 3 for 0.2 % DMSO.α3β2ε: n = 4 for GABA-control response, n = 3 for GABA + 0.1 % DMSO, n = 2 for 0.1 % DMSO,n = 3 for GABA + 0.2 % DMSO, n = 3 for 0.2 % DMSO. α3β2γ2: n = 9 for GABA-control response,n = 5 for GABA + 0.1 % DMSO, n = 3 for 0.1 % DMSO, n = 4 for GABA + 0.2 % DMSO, n = 5 for0.2 % DMSO. Data are means ± SEM; n represents the number of oocytes examined. Currents werenormalised to the maximal response induced by GABA at EC20 concentrations.

195

Chapter B. Additional Graphs

40

60

80

100

120

10 20 30

NaI

100 µM GABA

2.45 µM GABA

2.45 µM GABA + 0.1% DMSO





BF


2.45 µM GABA + 1% DMSO

Time [sec]

% net fluorescence

40

60

80

100

120

10 20 30

8.26 µM GABA + 1% DMSO







8.26 µM GABA

100 µM GABA

NaI

BF

Time [sec]

% net fluorescence

A

B

Figure B.2: Fluorescence signal of mutant YFP-H148Q/I152L in response to differentconcentrations of DMSO in HEK293 cells also expressing α3β2γ2- (A) or α1β2ε GABAAreceptors .Fluorescence was measured for 30 seconds after addition of NaI test solution only or together with100 µM GABA, EC30 GABA or EC30 GABA and 0.1 to 1 % DMSO. Net fluorescence was norma-lised by setting the averaged baseline fluorescence to 100 % and fluorescence changes (after addingsolutions) were calculated as percentage of the baseline fluorescence.

196


10 µµµµM GABA

10 se

c

20 se

c

30 se

c

45 se

c1 m

in2 m

in3 m

in5 m

in0

1

2

3

Time

OD

405

DPBS

10 se

c

20 se

c

30 se

c

45 se

c1

min

2 m

in3

min

5 m

in0

1

2

3

**** ***

***

** *****

**

**

Time

OD

405

A

B

Figure B.3: OD405 values from iodide-flux assay in HEK293 cells expressing α3β2γ2GABAA receptors. Before performing the SK assay, 10 µM GABA or DPBS were incu-bated for ten seconds to five minutes, respectively.Readings represent data from SK assay from undiluted cell lysate from assays, which were repeatedthree times (first assay = yellow columns, second = green columns, third = blue columns). Valuesare means ± SD from six data points each. * indicates p < 0.05, ** = p < 0.01, *** = p < 0.001comparing GABA-treated cells (A) with their respective DPBS controls (B).

197


0 60 120 180 240 3001.0

1.5

2.0

2.5GABADPBS

Time [sec]

OD

405

0 60 120 180 240 3000

1

2

3GABADPBS

Time [sec]

OD

405

0 60 120 180 240 3000

1

2

3GABADPBS

Time [sec]

OD

405

A

B

C

Figure B.4: OD405 values from iodide-flux assay in HEK293 cells expressing α3β2γ2GABAA receptors. Before performing the SK assay, 10 µM GABA or DPBS were incu-bated for ten seconds to five minutes, respectively.Readings represent data from SK assay from (A) undiluted cell lysate from assays (n = 18). (B) Valuesobtained from diluted cell supernatant, 2:3 (n = 6) or (C) 1:5 (n = 6). Values are means ± SEM.

198


-8 -7 -6 -5 -4 -3 -2

0

25

50

75

100 α1β2ε

log [GABA, M]

% m

axim

um

res

po

nse

Figure B.5: GABA concentration-response curve for α1β2ε GABAA receptors (n = 4) ex-pressed in Xenopus laevis oocytes. The data correspond to means ± SEM; n states thenumber of oocytes tested. The EC50 value was 3.553 ± 0.52 µM and the Hill coefficient0.71. Curves were fitted using a non-linear regression fit with variable slopes as described inSection 2.3.5.

-8 -7 -6 -5 -4 -3 -2

0

25

50

75

100 α3β2γ2

log [GABA, M]

% m

axim

um

res

po

nse

Figure B.6: GABA concentration-response curve for α3β2γ2 GABAA receptors (n = 10)expressed in HEK293 cells. The data correspond to means ± SEM; n states the number ofcells clamped. The EC50 value was 14.65 µM and the Hill coefficient 1.88. Curves werefitted using a non-linear regression fit with variable slopes as described in Section 2.3.5.

199

studies on the promiscuous subunit tina schwabeirep.ntu.ac.uk/id/eprint/197/1/203555_tinas...

Documents